How We Evaluate Sources and Conflicting Evidence

sherafy.com welcomes scrutiny.

If we publish something that is factually wrong, materially incomplete, based on weak evidence or contradicted by stronger evidence, we want to know.

But taking criticism seriously does not mean treating every rebuttal, article, organization, argument or collection of links as equally credible.

Our standard is simpler:

Every claim deserves fair scrutiny. Not every claim survives that scrutiny, and not every source therefore deserves equal evidentiary weight.

That principle applies across sherafy.com—from medicine and scientific research to politics, technology, corporate behavior, history, war and human rights.

It is especially important in our Israel-Palestine coverage, where factual disagreements are frequently entangled with identity, trauma, ideology, national loyalty, advocacy and competing historical narratives.

This page explains how we handle those conflicts.

Equal scrutiny does not mean equal weight

Objectivity is sometimes described as presenting both sides.

That can be useful when two competing interpretations are genuinely supported by substantial evidence.

But it is not a sufficient research method.

If one explanation is supported by authenticated records, contemporaneous evidence and multiple independent sources while another rests primarily on unsupported assertions, selective quotations or repetition, presenting them as equally established would not make an article more objective.

It would make the article less accurate.

Fairness concerns the standard applied to competing claims.

It does not require an equal outcome after that standard has been applied.

A conclusion supported by considerably stronger evidence should normally receive considerably greater evidentiary weight.

We do not adjust that weight simply to create the appearance of political symmetry.

We evaluate the evidence, not just the name on the source

Source reputation matters, but it is not the end of the inquiry.

A government can possess excellent primary information and still have incentives to characterize it favorably.

An advocacy organization can have an unmistakable political agenda and still uncover an authentic document that proves something important.

A highly respected newspaper can misunderstand a study.

An obscure researcher can locate a primary record everyone else missed.

A witness can provide essential firsthand information while still being capable of error.

A scientific journal can publish a legitimate study whose findings are subsequently exaggerated far beyond what the experiment established.

For those reasons, sherafy.com asks more than:

Who published this?

We also ask:

What does the underlying evidence actually establish?

Whenever practical, we try to move backward through the information chain toward the original material.

If an advocacy article cites a court filing, we want the filing.

If a news story cites a scientific study, we want the study.

If a political organization quotes a government document, we want the complete document.

If a viral post relies on a video, we want to establish the video’s origin, date, location and context.

If twenty articles repeat the same allegation, we want to determine whether they represent twenty independent confirmations or twenty repetitions of one original claim.

The intermediary can matter.

The underlying evidence usually matters more.

A source can be biased and still contain valid evidence

“Bias” is often treated as though it were a binary property.

It is not.

Sources have perspectives, interests and incentives to different degrees, and those characteristics can affect which facts they emphasize, what questions they ask and how they interpret ambiguous evidence.

That is worth considering.

But identifying bias does not determine whether a factual claim is true or false.

An interested party can provide authentic evidence.

A supposedly neutral source can make a factual error.

The appropriate response is therefore neither automatic trust nor automatic dismissal.

It is verification.

When an interested party provides information, we may use it to locate evidence, understand that party’s position or identify questions requiring investigation.

When the underlying factual material can be independently established, we evaluate that material on its own merits.

When it cannot be independently established, the source’s access, incentives, track record, methodology and other circumstances become more important to determining how confidently the claim can be reported.

We evaluate claims, not teams

sherafy.com does not maintain a permanent list of institutions whose statements are automatically true or automatically false.

That would be a poor substitute for research.

A source can be highly reliable on one question and weak on another.

A single report can even contain conclusions with different levels of evidentiary support.

We therefore try to evaluate consequential claims individually.

That means asking questions such as:

• What specific proposition is being asserted?
• What evidence is offered for it?
• Is that evidence relevant to the proposition?
• Can it be independently verified?
• Is the original material available?
• Has important context been omitted?
• Does the conclusion logically follow from the evidence?
• Are multiple supposedly independent sources actually tracing back to the same origin?
• Is contradictory evidence addressed fairly?
• What evidence would have to exist for the claim to be wrong?
• Would we evaluate the same evidence differently if the political identities involved were reversed?

The last question is particularly important.

A standard that changes depending on whose conduct is being investigated is not a meaningful evidentiary standard.

Why Israel-Palestine receives special attention on this page

The standard described here applies everywhere on sherafy.com.

Israel-Palestine nevertheless deserves explicit treatment.

Our Israel-Palestine reporting receives some of the site’s most intensive reader scrutiny and some of the strongest challenges over which sources should or should not be included.

That is understandable.

The subject involves extraordinary human suffering, competing national narratives, religious and historical identity, terrorism, occupation, war, displacement, antisemitism, Palestinian dispossession, international law, information warfare and political advocacy operating simultaneously.

Many readers also approach the subject with deeply held prior beliefs.

None of that makes careful research impossible.

It makes careful research more important.

Our goal is not to construct an “Israeli side” and a “Palestinian side” and mechanically allocate equal space to each.

Our goal is to determine what the available evidence supports about the particular question being investigated.

Israeli government and military claims

Statements from the Israeli government, Israel Defense Forces or other Israeli state institutions can provide important information, including information that outside investigators may not possess.

They are also statements from parties directly involved in many of the events being investigated.

We therefore do not automatically accept or reject them.

Where consequential claims can reasonably be checked against photographs, video, satellite imagery, documents, witness testimony, forensic evidence, independent reporting or subsequent investigations, we try to perform that comparison.

An official statement remains evidence.

Official status does not make independent verification unnecessary.

Palestinian authorities, armed groups and other interested parties

The same principle applies to Hamas, the Palestinian Authority, armed organizations, political movements and other parties with direct interests in the conflict.

Their statements can contain important information.

They can also contain advocacy, incomplete information, strategic messaging or claims that require corroboration.

The political identity of the speaker does not alter the standard.

United Nations bodies, NGOs and human-rights organizations

We do not treat the name of an institution as a substitute for examining its work.

When a UN body, humanitarian organization, human-rights organization or investigative NGO publishes a consequential finding, relevant questions can include:

• What methodology was used?
• What primary evidence was reviewed?
• How were witnesses identified or corroborated?
• What limitations did investigators acknowledge?
• Were investigators able to access the location?
• Are underlying documents or datasets available?
• What exactly does the report conclude?
• Does a news headline go further than the report itself?

Institutional credibility can be relevant.

Methodology and evidence remain essential.

Israeli, Palestinian and international media

A publication is not accepted or rejected merely because it is Israeli, Palestinian, Arab, American, European or associated with a particular political tendency.

Original reporting can be valuable regardless of where it originates.

We look at what reporters actually established, the transparency of their sourcing, whether material claims are corroborated and whether primary evidence can be examined.

Images, videos and social-media evidence

Visual evidence can be powerful and also unusually vulnerable to misidentification.

When an image or video materially affects a conclusion, relevant questions can include provenance, date, location, original uploader, edits, cropping, sequencing and whether the material is being described accurately.

A real video presented with a false location can support a false conclusion.

Authentication and context therefore matter together.

Casualty figures and humanitarian statistics

Conflict statistics often involve incomplete access, changing classifications, retrospective identification, duplicated or corrected records and disagreements over terminology.

Where those limitations matter, we try to explain them.

A dataset should not be represented as more precise than its methodology permits.

Neither should uncertainty be exaggerated into the claim that nothing meaningful can be known.

Legal claims

Questions involving occupation, war crimes, genocide, ethnic cleansing, proportionality, collective punishment or other legal concepts require particular care.

Political use of a legal term is not identical to a judicial determination.

We therefore try to distinguish among:

• treaty language;
• statutes and legal instruments;
• binding court judgments;
• provisional measures;
• arrest warrants;
• advisory opinions;
• commission or expert findings;
• allegations by parties;
• scholarly legal interpretation; and
• ordinary moral or political use of legal terminology.

Those categories can overlap.

They are not interchangeable.

Historical claims

Historical narratives can become political weapons precisely because modern arguments often depend on what people believe happened decades earlier.

Where practical, we prefer archival records, contemporaneous documents, demographic data, original statements and serious historical scholarship over simplified retrospective narratives.

Historical disagreement should be presented where it genuinely exists.

The existence of disagreement does not mean every historical claim is equally plausible.

Reader-submitted research is welcome

Readers regularly send us articles, documents, studies, videos, rebuttals and arguments that they believe change the picture presented in an article.

We encourage that.

Contrary evidence is useful.

In fact, one of the strongest ways to test a conclusion is to actively look for evidence capable of disproving it.

But material does not become credible merely because it contradicts something we published.

A rebuttal deserves examination because it might be right.

It does not deserve publication merely because it exists.

If submitted material identifies a genuine factual error, overlooked primary document, missing context, stronger dataset, misidentified photograph, incorrect quotation or other evidence materially affecting an article, we should address it.

That can mean clarification.

It can mean adding evidence.

It can mean changing a conclusion.

And when necessary, it can mean issuing a correction.

What we will not do is increase the apparent credibility of a weak claim merely because someone demands that it receive “equal time.”

What makes a useful correction or challenge

The most useful criticism identifies something specific.

For example:

“This sentence is factually incorrect.”

Then provide the evidence demonstrating the error.

Or:

“This quotation omits context that materially changes its meaning.”

Then provide the complete source.

Or:

“You describe these as three independent confirmations, but all three ultimately rely on the same original witness.”

Show the sourcing chain.

Or:

“This dataset was revised after publication and the current numbers are different.”

Provide the updated dataset.

Or:

“There is a primary document you did not include that materially weakens your conclusion.”

Provide it and explain what it establishes.

Those challenges are valuable whether they ultimately prove us wrong or strengthen the original conclusion.

What does not become evidence through repetition

Research quality is not measured by the number of links in a message.

Ten pages repeating one unsupported assertion do not create ten independent pieces of evidence.

A hundred social-media posts do not authenticate a photograph.

Multiple articles citing one another in a circle do not create corroboration.

A very long argument does not become strong merely because responding to every individual assertion would require substantial time.

The relevant questions remain:

What is the claim?

What evidence supports it?

Where did that evidence originate?

How strong is it?

Rhetoric does not repair weak evidence

Public disputes frequently contain arguments about motives, hypocrisy, identity, institutional reputation and the character of the people involved.

Some of those issues may be relevant context.

They do not substitute for evidence establishing the disputed proposition.

Attacking the journalist who found a document does not make the document disappear.

Calling an organization biased does not establish that a particular statistic is false.

Pointing to wrongdoing by the other side does not establish whether the event currently being investigated happened.

Invoking someone’s religion, nationality or political affiliation does not resolve a forensic question.

And describing a conclusion as offensive, dangerous, antisemitic, anti-Palestinian, partisan, establishment, conspiratorial or politically inconvenient does not by itself establish whether the underlying factual claim is correct.

Those are separate questions.

When criticism identifies a genuine problem with evidence, we should address that problem.

When it does not, rhetoric cannot be used to manufacture an evidentiary dispute that the underlying record does not support.

Context matters—but “more context” has to change something

One of the most legitimate criticisms of journalism is that technically accurate facts can still mislead when important context is omitted.

We take that seriously.

But context has to be material.

The relevant question is not whether additional information exists. Additional information almost always exists.

The question is:

Would including this information materially change how a reasonable reader understands the claim being evaluated?

If yes, it belongs in the analysis.

If no, the mere existence of additional facts does not demonstrate that the original reporting was misleading.

This distinction prevents both selective omission and the opposite tactic: burying a straightforward factual question beneath large quantities of tangential information until the original question disappears.

We do not judge research by ideological usefulness

Evidence supporting a conclusion we already favor does not receive a discount on scrutiny.

If a source supporting our argument uses a false quotation, we should not use it.

If a pro-Palestinian claim depends on a misidentified photograph, the photograph should be rejected.

If an antiwar argument relies on a statistic that cannot be substantiated, the statistic should not be repeated.

If a consumer advocate exaggerates a scientific study, the exaggeration should be corrected.

If criticism of a corporation rests on documents that do not actually say what critics claim, that matters.

If evidence favorable to an institution we have previously criticized disproves an allegation, that evidence matters too.

The standard only means something if we are willing to lose an argument because the evidence changed.

Strong sources can still be wrong

Evidence hierarchy is useful, but it should never become credential worship.

Primary documents can contain false statements.

Government statistics can contain methodological limitations.

Peer-reviewed studies can fail replication.

Courts can overturn earlier decisions.

Eyewitnesses can misremember events.

Journalists can make mistakes.

Experts can disagree.

Institutions can correct their own previous findings.

That is why sherafy.com does not ask readers to trust a conclusion merely because an authoritative source appears in the bibliography.

The purpose of sourcing is to make verification easier.

Weak sources can sometimes lead to strong evidence

The opposite is also true.

A partisan blog, anonymous account, advocacy organization or obscure social-media user may occasionally identify an authentic record that other researchers overlooked.

We do not need to adopt the source’s worldview to examine what it found.

If the underlying document is authentic, relevant and independently verifiable, we can use the evidence itself.

This is one reason we prefer tracing claims backward rather than merely maintaining lists of approved and disapproved publications.

Discovery and verification are different functions.

False balance can be a form of misinformation

Suppose one explanation is supported by extensive authenticated evidence and another has little credible support.

Giving each one half of an article may look impartial.

It can also leave readers with a profoundly inaccurate understanding of the state of the evidence.

The job of research is not to manufacture a tie.

Sometimes the evidence is genuinely uncertain.

We should say so.

Sometimes multiple explanations remain plausible.

We should explain them.

And sometimes one conclusion is overwhelmingly better supported.

We should be willing to say that too.

Objectivity requires equality of scrutiny, not equality of evidentiary weight.

What would make us change our conclusion?

This is one of the most important questions in serious research.

A conclusion that cannot theoretically be disproved is not being treated as an evidence-based conclusion.

Depending on the subject, evidence capable of changing our assessment might include:

• a previously unavailable primary record;
• authenticated evidence contradicting a key factual premise;
• a materially better dataset;
• proof that evidence was fabricated or misidentified;
• a credible alternative explanation accounting for the same facts more successfully;
• a corrected scientific analysis;
• newly released archival material;
• subsequent legal findings;
• independently corroborated testimony; or
• evidence demonstrating that we misunderstood an important source.

We cannot promise that every disagreement will end with everyone reaching the same conclusion.

We can promise that disagreement alone will not protect one.

Corrections are not defeats

Research is not strengthened by defending an obsolete conclusion.

If credible new evidence establishes that something we published is materially wrong, the appropriate response is to correct it.

That principle applies particularly to fast-developing conflicts such as Israel-Palestine, where casualty information, investigations, legal proceedings, official accounts and independently verified evidence can change after an article is originally published.

It applies equally to medicine, where new trials can overturn earlier expectations; technology, where specifications and policies change; legal reporting, where later decisions alter the governing record; and every other subject we cover.

Accuracy is the objective.

Being permanently consistent with ourselves is not.

The standard we ask readers to use on us

Nothing on this page means sherafy.com should receive special deference.

Readers should scrutinize our work too.

Check whether the sources support what we say they support.

Check whether quotations preserve their meaning.

Check whether evidence is being characterized accurately.

Check whether facts contradicting our conclusion have been ignored.

Check whether inference is being presented as fact.

Check whether newer evidence has superseded what we used.

Check our reasoning.

And if you find a material problem, bring us the evidence.

We would rather correct an article than preserve an error.

The principle in one sentence

Our source policy can ultimately be reduced to this:

Bring us contrary evidence. We will examine it seriously. But fairness requires that every claim face the same standard of proof—not that strong and weak evidence be made to look equal.

That standard applies to Israel and Palestine.

It applies to governments and their critics.

It applies to corporations and consumers.

It applies to scientific institutions and people challenging them.

It applies to conventional explanations and unconventional ones.

And it applies to sherafy.com itself.