Source Synthesis | Complete Guide to Combining Multiple Sources, Agreements, Disagreements and Evidence

Writing & Composition · Academic English · Multilingual English & Mediation

Source synthesis combines information from multiple sources around a shared question, concept or claim. Strong synthesis is not a stack of summaries. It shows how sources agree, differ, qualify one another, use different methods or leave gaps that matter to the writer’s purpose.

This SETC guide explains synthesis questions, source matrices, themes, agreement, disagreement, qualification, method differences, source integration, paragraph design, comparative language, citation, avoiding patchwriting and examination synthesis.

The governing rule is organise by idea, not by author. Sources should meet inside the writer’s structure rather than occupying isolated paragraphs simply because they were read separately.

What is Source Synthesis?

Source synthesis is the process of integrating evidence and ideas from two or more sources in order to explain relationships, evaluate patterns or build a larger answer that no single source provides alone.

Source Synthesis is best treated as a reader-facing system rather than a surface format. The writer decides what the task requires, what evidence belongs, how the material should be organised and what level of explanation lets a reader recover the intended meaning without guesswork.

Why source synthesis matters

Most substantial research questions cannot be answered reliably from one source. Synthesis allows writers to compare evidence, identify consensus, recognise disagreement and construct a more complete model of the subject.

The long-term goal is transfer. A student who can produce source synthesis only from a familiar model has learned an example. A student who can adapt the same underlying decisions to a new topic, audience, source set and time limit has learned the skill.

What source synthesis is not

one paragraph per source
copying several quotations into the same paragraph
using multiple citations after a claim without explaining the relationship
forcing sources to agree
treating disagreement as proof that nobody knows anything
losing source boundaries in blended paraphrase

These boundaries matter because a piece can look formal while remaining weak. Headings, citations, technical vocabulary and long sentences do not compensate for unclear purpose, poor source use or unsupported claims.

The operating sequence

Clarify the task, audience, scope and required outcome.
Gather enough information or source material to permit selection.
Define the central question, thesis or reporting purpose.
Group material by function rather than by the order in which it was found.
Choose a structure that matches reader needs.
Draft the major sections before polishing local wording.
Revise for evidence, coherence, source boundaries and clarity.
Proofread and check citation or formatting requirements last.

This sequence reduces wasted effort. Writers often lose time polishing an early paragraph before they know whether the evidence, structure or purpose will survive revision. Larger decisions should stabilise first.

The synthesis question

A clear question determines which relationships among sources matter and which differences are irrelevant.

For diagnosis, ask what job the synthesis question performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle the synthesis question differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

For transfer, change the topic, audience, source type or time limit while preserving the underlying writing problem. Mastery appears when the learner can rebuild the quality without copying the earlier wording or section order.

Source matrix

A matrix can organise sources against themes, findings, methods or criteria so patterns become visible before drafting.

For diagnosis, ask what job source matrix performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle source matrix differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Agreement

Sources may converge on a finding even when they use different evidence or terminology.

For diagnosis, ask what job agreement performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle agreement differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Partial agreement

Two sources may agree on a broad pattern but differ on magnitude, cause or interpretation.

For diagnosis, ask what job partial agreement performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle partial agreement differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Disagreement

Conflicting claims should be analysed rather than merely placed side by side.

For diagnosis, ask what job disagreement performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle disagreement differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Method differences

Different samples, definitions, measurements or timeframes can explain apparent contradiction.

For diagnosis, ask what job method differences performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle method differences differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Qualification

One source may narrow or condition a claim made more broadly by another.

For diagnosis, ask what job qualification performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle qualification differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Extension

A source can extend another by adding a new population, mechanism, example or implication.

For diagnosis, ask what job extension performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle extension differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Chronology

Later research may confirm, revise or challenge earlier findings, making time an important synthesis dimension.

For diagnosis, ask what job chronology performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle chronology differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Theme organisation

Paragraphs often work best when organised around themes or questions rather than source names.

For diagnosis, ask what job theme organisation performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle theme organisation differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Claim-first paragraphs

A synthesis paragraph should normally begin with the writer’s analytical claim before sources are introduced as evidence.

For diagnosis, ask what job claim-first paragraphs performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle claim-first paragraphs differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Comparative language

Phrases such as similarly, in contrast, however and whereas should correspond to real source relationships.

For diagnosis, ask what job comparative language performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle comparative language differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Citation clusters

Multiple sources can support one sentence or paragraph when attribution remains clear enough to identify who supports what.

For diagnosis, ask what job citation clusters performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle citation clusters differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Source boundaries

The writer must keep each source’s claims, wording and limitations distinguishable even while combining them.

For diagnosis, ask what job source boundaries performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle source boundaries differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Paraphrase control

Synthesis should not become patchwriting in which fragments of several source sentences are stitched together.

For diagnosis, ask what job paraphrase control performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle paraphrase control differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Counterevidence

A synthesis is stronger when it includes relevant evidence that complicates the emerging pattern.

For diagnosis, ask what job counterevidence performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle counterevidence differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Gaps

Missing populations, methods or questions can become part of the synthesis when they limit what the literature can establish.

For diagnosis, ask what job gaps performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle gaps differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

Exam synthesis

Timed source-based writing should group sources by the question or criterion instead of summarising each text in turn.

For diagnosis, ask what job exam synthesis performs for the reader. If the writer cannot answer that question, the section may exist because a template suggested it rather than because the document needs it. The repair begins by restoring purpose.

For deliberate practice, create two short versions that handle exam synthesis differently. Compare them using explicit criteria such as evidence quality, reader orientation, transparency, logic, accessibility or efficiency. This turns revision into a decision process rather than a matter of taste.

A diagnostic checklist

Is there a clear synthesis question?
Are sources organised by idea rather than reading order?
Can agreements and disagreements be stated precisely?
Are method differences considered?
Does each paragraph contain the writer’s own analytical claim?
Are source boundaries clear?
Is counterevidence included where relevant?
Does the synthesis produce an understanding larger than any one source summary?

Use the checklist to identify the first weak link rather than correcting every visible symptom. One early problem—such as a vague research question, unclear reporting purpose or weak source boundary—can create several later errors at once.

Worked example: Maren

Maren writes four paragraphs beginning Smith says, Jones says, Lee says and Tan says. Her tutor asks what question all four sources are helping answer. She reorganises by two themes, allowing several authors to interact inside each paragraph.

Maren is asked to explain the reader need, revise one section and state why the second version is stronger. The explanation converts feedback into a reusable rule instead of a one-time correction.

Worked example: Iona

Iona sees two sources disagree and writes that the evidence is inconclusive. Her tutor compares the studies’ populations and finds that they examined different age groups. The apparent contradiction becomes a useful boundary rather than a dead end.

Iona is asked to explain the reader need, revise one section and state why the second version is stronger. The explanation converts feedback into a reusable rule instead of a one-time correction.

Worked example: Leonie

Leonie spends too much exam time describing every source. She creates a quick matrix of agree, disagree and extend, then writes paragraphs around those relationships rather than around source order.

Leonie is asked to explain the reader need, revise one section and state why the second version is stronger. The explanation converts feedback into a reusable rule instead of a one-time correction.

Planning before drafting

Planning should identify the central purpose, major sections and evidence responsibilities without becoming a second full draft. Its value lies in reducing structural uncertainty before sentence-level writing begins.

A practical mastery check is to apply this principle to a fresh task, explain the decision aloud, revise once and repeat the same move after a delay without the earlier model. This tests retrieval, adaptation and independent control.

Gathering material

Writers need more material than the final document will use. Research notes, examples, observations and source excerpts should be collected with enough context that they can later be evaluated accurately.

Selection

Not everything discovered deserves inclusion. Material should survive because it is relevant, credible and useful to the reader’s task, not because it was difficult to find.

Hierarchy

A strong document distinguishes central claims or findings from supporting detail. Space, headings and paragraph development should reflect importance.

Organisation

Order may follow argument, chronology, method, results, themes, criteria or reader questions. The writer should choose deliberately rather than inherit the order of notes.

Paragraph roles

Each paragraph should perform one local job. Clear paragraph roles make evidence placement, transitions and revision easier.

Evidence placement

Evidence should appear close enough to the claim or finding it supports that the relationship remains visible. Delayed explanation forces the reader to carry unresolved material.

Source boundaries

Writers should distinguish source ideas, source wording, their own analysis and established background knowledge. Clear attribution prevents both plagiarism and conceptual confusion.

Paraphrase and quotation

Paraphrase should genuinely restate source meaning, while quotation should be reserved for wording worth preserving exactly. Both require appropriate attribution.

Synthesis

Synthesis organises multiple sources around ideas, questions or findings rather than summarising one source at a time. It reveals agreement, disagreement, extension and gaps.

Transitions

Transitions should reflect real relationships such as contrast, extension, cause, method or implication. They should not hide weak organisation.

Clarity and register

Academic and professional writing should be precise, readable and appropriately formal. Complexity should come from the subject, not from unnecessary verbal difficulty.

Revision with a reverse outline

After drafting, label what each section actually does. A reverse outline exposes repetition, gaps, misplaced evidence and sections whose real role differs from the plan.

Cutting and adding

Revision may remove duplicated background or weak examples while adding a missing definition, source comparison, limitation or explanatory bridge. The aim is functional completeness rather than maximum length.

Feedback and self-correction

Feedback becomes durable when translated into a process rule. A learner should leave with a principle that can be applied to the next document.

Spacing and delayed practice

Delayed practice tests whether the writer can reconstruct the process without a model active in working memory. It is a stronger mastery check than immediate repetition.

Cross-curricular transfer

Source Synthesis supports school, university and workplace communication. The conventions vary, but writers still need controlled purpose, evidence, structure and reader awareness.

Primary and PSLE application

Primary learners can begin by comparing two short sources and identifying one shared idea and one difference with clear attribution.

Secondary and SEC application

Secondary writers should organise multiple sources by themes, compare evidence and preserve attribution rather than writing source-by-source summaries.

JC, IGCSE and advanced application

Advanced writers should analyse methods, definitions, populations, chronology, counterevidence and research gaps while building coherent synthesis paragraphs.

Exam and timed use

Under time pressure, protect task fulfilment, evidence, major structure and completion before small stylistic polish. High-value decisions should receive time first.

Measuring progress

Improvement may appear as faster planning, cleaner source boundaries, stronger paragraph roles, more accurate evidence use and better completion before it appears as a large score change.

Frequently asked questions

Is formal language the same as difficult language?

No. Strong formal writing is precise and controlled, not unnecessarily complicated.

How many sources do I need?

Enough high-quality sources to support the task and scope. Source count alone does not guarantee quality.

Should every source get its own paragraph?

Usually no. Organise around ideas, findings or questions unless the task specifically requires source-by-source treatment.

When should I quote instead of paraphrase?

Quote when the exact wording matters; paraphrase when the idea matters more than the original phrasing.

What should I revise first?

Start with purpose, evidence, structure and missing explanation before local style.

How do I improve under time pressure?

Practise a compact sequence of planning, evidence selection and high-value revision until it becomes reliable.

Research and reference points

Purdue OWL’s research and synthesis guidance emphasises integrating sources into the writer’s own argument and maintaining clear attribution. Current academic-writing guidance from Grammarly likewise distinguishes research structure and citation from simple source accumulation.

Continue through SETC

Continue with Research Writing, Paraphrasing and English Writing and Composition.

The mastery target is independent control: recognise when source synthesis is needed, choose an appropriate structure, use evidence transparently and revise the document for an unfamiliar reader without relying on one fixed template.

When Two Sources Disagree: Write a Synthesis Without Inventing Consensus

Disagreement is not a defect that the writer must remove. It is information. When two credible sources reach different conclusions, the writer’s job is not to force them into a single sentence that sounds tidy. The job is to preserve the shape of the disagreement, identify what each source actually claims, examine why the claims differ, and write a synthesis that lets the reader see both the common ground and the unresolved difference.

This matters because weak synthesis often becomes false consensus. A student reads two sources, notices that both discuss the same topic, and writes that “both authors agree” even when they agree only on the problem, not on the cause, scale, solution or certainty. Another student places one source after the other and calls the paragraph a synthesis. A third chooses one source, dismisses the other and quietly converts a comparison task into an argument for a preferred view. None of these moves solves the actual writing problem.

The stronger method is slower at the beginning and faster at the end. Build an accurate model of the relationship before drafting. Once the relationship is clear, the paragraph becomes much easier to write.

The central question: exactly what is in conflict?

“The sources disagree” is too broad. Two sources can disagree in several different ways. They may disagree about whether an effect exists, how large it is, what caused it, how long it lasts, which population it applies to, which evidence should count, what intervention should be used, or how confident anyone should be. They may also appear to disagree because they are answering different questions.

Before writing, complete this sentence: Source A and Source B differ specifically about ______. If the blank cannot be filled with a precise proposition, the writer is not ready to synthesise.

Existence: one source reports an effect; the other does not.
Magnitude: both report an effect, but one describes it as large and the other as modest.
Cause: both observe the same pattern but explain it differently.
Scope: one makes a broad claim; the other limits the claim to a particular group or condition.
Method: the sources use different measurements, samples or definitions and therefore produce results that are not directly interchangeable.
Recommendation: the evidence may be similar while the proposed action differs.
Certainty: the sources may point in the same direction while expressing different levels of confidence.

These distinctions protect the writer from the temptation to turn every difference into a dramatic contradiction. Sometimes the accurate relationship is partial agreement with a narrower disagreement inside it.

An original two-source practice set

The following two passages are original SETC practice material created for this lab. They are not quotations from published research. Their purpose is to make the synthesis decisions visible before the learner transfers the method to real sources.

Practice Source A

A school introduces a twenty-minute independent reading period at the start of three English lessons each week. After one term, teachers report that many students begin longer reading tasks more readily and require fewer reminders to stay on task. The school also records a small improvement in average reading-comprehension scores. The programme coordinator argues that protected reading time helped students build reading stamina and made extended reading feel more normal. She cautions that the school did not randomly assign students to different conditions and that several teachers also changed their classroom routines during the same term.

Practice Source B

A second school runs a similar reading period but finds little change in comprehension scores across the term. Student interviews show that some pupils enjoyed the quiet routine, while others chose books they could finish easily and rarely attempted more demanding texts. The research lead concludes that independent reading time may support reading habits, but time alone is unlikely to improve comprehension unless students also receive guidance in selecting texts, discussing meaning and repairing misunderstandings. He notes that the school’s assessment measured inference and vocabulary more heavily than reading speed or willingness to read.

Do not start by deciding who is right

A novice response often asks which source should “win”. That is not the first synthesis question. Start by mapping the claims.

Issue	Source A	Source B	Relationship
Reading routine	Protected time appeared to improve readiness and persistence.	Some students valued the routine.	Partial agreement.
Comprehension	Small average improvement was recorded.	Little change was recorded.	Different findings.
Mechanism	Stamina and normalisation of extended reading may matter.	Text choice, discussion and repair support may be necessary.	Different emphasis; not mutually exclusive.
Limitations	No random assignment; other classroom changes occurred.	Different assessment emphasis; students selected easy texts.	Both qualify their conclusions.
Practical implication	Protect regular reading time.	Pair reading time with guidance and meaning-focused support.	Source B narrows the conditions under which the routine may help.

Now the relationship is more precise. The sources do not simply “agree” or “disagree”. They share a view that routine may matter, report different comprehension outcomes, and differ in how much additional support they think is necessary. A strong synthesis paragraph should preserve that layered relationship.

The false-consensus paragraph

Weak version: Both sources agree that independent reading time improves students’ reading and should be used in schools. Source A shows that reading scores improved, while Source B says that students enjoyed the routine. Therefore, regular independent reading is an effective way to improve comprehension.

This paragraph is smooth and inaccurate. It commits several errors at once. It converts Source B’s cautious claim into agreement that comprehension improves. It ignores Source B’s finding of little change in comprehension scores. It treats enjoyment as proof of comprehension. It removes Source A’s design limitations. Finally, it upgrades a qualified relationship into a confident recommendation.

The paragraph sounds more decisive than the evidence. That is exactly why it is dangerous.

The source-by-source paragraph

Incomplete version: Source A studied a school with independent reading time and found a small improvement in comprehension. The coordinator thought reading stamina improved. However, there was no random assignment. Source B studied another school and found little improvement. Some students enjoyed the routine but chose easy books. The research lead said students also need guidance.

This version is more accurate, but it is still mostly two summaries placed next to each other. The writer has not told the reader what the relationship means. Synthesis begins when the sources are organised around the shared question rather than around the order in which they were read.

A stronger synthesis paragraph

Stronger version: The two school accounts suggest that protected reading time may support reading habits more reliably than it supports comprehension by itself. Source A recorded stronger task persistence and a small rise in comprehension scores, but its coordinator also noted that other classroom changes occurred during the same term. Source B found little change in comprehension and argued that students may need support with text selection, discussion and misunderstanding repair. Taken together, the sources do not establish that reading time alone raises comprehension. They instead point to a narrower possibility: a regular reading routine may create useful conditions for reading, while the quality of text choice and meaning-focused support may influence whether those conditions lead to deeper comprehension.

Notice what this paragraph does not do. It does not invent a single shared conclusion. It does not pretend the findings are identical. It does not choose a winner before examining the differences. It builds the most defensible relationship that the two sources can jointly support.

The four-layer synthesis test

When two sources disagree, test the paragraph across four layers.

Claim layer: What does each source actually claim?
Evidence layer: What evidence or method supports that claim?
Boundary layer: What population, condition, timeframe or limitation restricts the claim?
Relationship layer: Does the second source confirm, narrow, challenge, explain or simply address a different part of the first?

A paragraph that skips the boundary layer tends to overgeneralise. A paragraph that skips the evidence layer tends to treat all statements as equally supported. A paragraph that skips the relationship layer becomes serial summary.

Language for preserving disagreement

Writers need language that is precise enough to show relationships without exaggerating them. The following patterns are useful when they match the evidence.

Both sources identify ___, but they differ on ___.
Source A reports ___, whereas Source B finds ___ under a different condition.
Source B does not reject Source A’s observation; it narrows the conditions under which the observation may apply.
The apparent contradiction may partly reflect differences in ___.
The sources converge on ___ while remaining divided over ___.
Source A attributes the pattern mainly to ___; Source B gives greater weight to ___.
Taken together, the evidence supports ___ more clearly than it supports ___.
The sources do not establish a single conclusion about ___, but they jointly show ___.
Source B qualifies the broader claim by showing ___.
The disagreement remains unresolved because the sources use different ___.

These are not templates to paste automatically. Each phrase makes a logical claim about the relationship between sources. The writer must be able to defend that claim.

Method differences can create apparent contradiction

Two sources may seem to disagree when their studies are not directly comparable. Check the following before treating a difference as a conflict:

Were the same outcomes measured?
Were the participants similar?
Did the studies run for the same length of time?
Did the sources define the key concept in the same way?
Were the conditions equally controlled?
Were the same comparison groups used?
Was one source reporting self-perception while the other measured performance?
Was one source discussing immediate effects and the other long-term effects?

If the answer is no, the synthesis should name the mismatch. “The studies reached different results” may be technically true but analytically weak if the studies measured different things.

Do not erase minority evidence

Consensus is not created by counting citations casually. Five sources making one claim and two making another does not automatically mean the five are correct. Source quality, method, independence, relevance and recency may matter. Several sources may also repeat the same underlying evidence rather than provide five independent confirmations.

A careful synthesis can report a dominant pattern while preserving a credible exception. For example: “Most of the reviewed studies report a small positive association, although the two studies using direct performance measures rather than self-report do not show the same pattern.” That sentence tells the reader both the apparent majority and the methodological complication.

Do not convert uncertainty into neutrality

Preserving disagreement does not mean pretending every position is equally supported. If one source provides strong evidence and another offers a weakly supported opinion, the synthesis should say so. Neutrality is not the same as accuracy. The writer can compare evidence quality without becoming polemical.

Useful questions include: Is the source describing data, interpreting data, offering an expert judgement or making a recommendation? Does it disclose limitations? Is the evidence directly relevant to the claim? Does the conclusion exceed the design? Does another source independently support the same point?

A practical drafting sequence

Write the shared question at the top of the page.
Reduce each source to one or two precise propositions.
Mark words that carry scope or certainty: may, most, only, under, among, during, associated with.
List the evidence or method behind each proposition.
Identify the exact point of agreement.
Identify the exact point of disagreement.
Ask whether method, population, definition or timeframe explains part of the difference.
Choose the relationship the paragraph must communicate.
Draft the relationship sentence before the source-detail sentences.
Add evidence from both sources.
End with what the combined evidence supports and what remains unresolved.
Reopen both sources and audit the paragraph for strengthened, weakened or invented claims.

The disagreement matrix

A simple matrix can prevent the writer from losing the structure.

Field	Source A	Source B
Main proposition
Evidence type
Population / context
Key limitation
Certainty language
What agrees
What differs
Possible reason for difference
What can safely be concluded
What cannot safely be concluded

Failure mode: averaging the claims

Students sometimes create a middle position that neither source states. If Source A says a programme produced a large effect and Source B says it produced no reliable effect, the synthesis cannot simply declare that the programme “probably has a moderate effect” unless there is evidence for that inference. Averaging words is not synthesis.

Failure mode: using “however” as analysis

Placing “however” between two sentences signals contrast but does not explain it. The reader still needs to know what differs and why the difference matters. Replace connector-only synthesis with relationship synthesis.

Failure mode: confusing topic overlap with agreement

Two sources can both discuss social media, sleep, reading or homework while making incompatible claims. Shared topic is not shared conclusion. Always compare propositions.

Failure mode: making one source answer the other source’s question

One source may ask whether an intervention changes scores. Another may ask how students experience the intervention. Their findings may be complementary rather than contradictory. A synthesis should not force them into the same evidential role.

Failure mode: hiding the disagreement in vague wording

Phrases such as “research is mixed” can be useful summaries only after the writer explains what is mixed. Otherwise they become an escape route. Name the variable, outcome or condition that produces the difference.

Feedback: mark the relationship, not only the grammar

When reviewing a synthesis paragraph, use four feedback questions before sentence-level editing.

Can the reader identify the exact shared question?
Can the reader see the exact point of agreement and disagreement?
Has either source been made more certain, broader or more causal than it was?
Does the final sentence state only what the combined evidence can support?

Grammar and style matter, but a grammatically polished false consensus is still a failed synthesis.

Fresh-source check

Transfer requires a new pair of sources. Do not retest the learner using the same practice passages with different questions. Use a fresh topic and repeat the decision process.

For example, create two original short reports about homework. Source C might report that structured homework improves retrieval when tasks are short and checked the next day. Source D might report that large homework loads are associated with sleep loss and lower completion quality. The learner must decide whether these sources truly disagree. A careful answer may conclude that they address different quantities and conditions rather than opposite principles.

The fresh check should require the learner to produce: a one-sentence relationship statement; a completed mini-matrix; one synthesis paragraph; one sentence naming what cannot yet be concluded; and a short explanation of why the paragraph did not force consensus.

Self-check before submission

I can state each source’s main proposition without looking at my paragraph.
I preserved important qualifiers and conditions.
I did not call topic overlap “agreement”.
I did not treat different measures as directly equivalent.
I explained the relationship rather than only using a connector.
I did not average incompatible claims into a new unsupported claim.
I distinguished evidence from interpretation.
I named unresolved uncertainty when it remained unresolved.
My concluding synthesis sentence is no stronger than the combined evidence.
I can perform the same process with a fresh source pair.

Where this skill transfers

The same mechanism appears in university essays, literature reviews, General Paper, source-based writing, reports, policy briefs and research notes. It also appears outside formal assessment whenever a reader must compare news reports, expert explanations, product claims or competing interpretations.

The durable skill is not “write a paragraph with two citations”. It is preserve relationships between claims while transforming multiple sources into one coherent answer.

Next routes

Continue with Writing the University Essay when the synthesis must support a thesis, How to Read a Research Paper as a Student when the challenge is interpreting research evidence accurately, and Academic English for the wider system of reading, evidence, source use and formal communication.