Evaluate a website before citing it in an assignment

A website can be useful academic evidence, but “it is online” tells you very little about what it can establish. A government data page, a charity campaign, a company report, a university explainer and an unsigned blog post may all be websites. They have different purposes, evidence bases and suitable uses.

The useful question is not simply “Is this a good website?” Ask instead:

Can this particular page support this particular assignment claim, and with what limits?

Answering that question requires more than inspecting the page itself. You need to identify who is responsible for it, trace important claims to their original evidence, compare independent coverage, check the date of the evidence and test relevance to the assignment. The result should be a bounded decision: use, use with limits or exclude.

This guide shows how to record that reasoning in a website evidence card. It does not ban web sources, treat a domain suffix as proof of quality or turn missing corroboration into an automatic accusation of falsehood.

State the assignment claim before judging the page

Evaluation becomes vague when the intended use remains “background” or “information about the topic”. Write the sentence or proposition the page might support.

Compare these possible uses:

• Remote mentoring programmes exist in higher education.
• Remote mentoring improves first-year retention.
• Students describe remote mentoring as increasing their sense of belonging.
• LearningBridge Institute publicly promotes remote mentoring as a retention intervention.

These claims need different evidence. An organisation’s own product page may establish what it publicly claims. It may not establish that the claimed outcome occurred. A survey of participants’ intentions may describe those intentions, but it does not by itself demonstrate later enrolment or a causal effect.

Also check the assignment’s source requirements. A page can be accurate and relevant yet still fail a requirement for peer-reviewed empirical research, primary historical material or evidence from a specified jurisdiction. Course rules are part of relevance, not an afterthought.

If you have not yet found candidate material, begin with the guide to building an assignment-scale search string. Website evaluation begins after a page has been selected for possible use.

Read laterally before reading the page more closely

Close reading asks what the page says. Lateral reading temporarily leaves the page to ask what independent sources say about the publisher, evidence or claim.

Ohio State University’s lateral-reading guidance distinguishes this from reading only within a site. The University of Southern California Libraries similarly recommends investigating the source, finding better coverage and tracing claims to their original context in its source-evaluation guide.

A practical first pass is:

1. Open a new search tab rather than relying on the site’s About page.
2. Search the organisation’s exact name and the names of any responsible authors.
3. Look for independent descriptions, relevant expertise, ownership and material interests.
4. Search a distinctive claim or statistic in quotation marks.
5. Follow links to the original report, dataset, policy or study.
6. Check whether credible coverage represents the evidence in the same way.

Lateral reading is not a popularity contest. Repetition across copied press releases is not independent corroboration. Silence also does not prove that a claim is false. Record what you found and what remains unresolved.

Build a website evidence card

A reusable evidence card keeps observations separate from the conclusions you draw from them.

Field	What to record	Why it matters
Intended claim	The exact assignment sentence or proposition the page may support	A source is suitable for a job, not credible in the abstract
Responsible party	Named author, responsible organisation, owner and relevant expertise	Responsibility helps you assess authority and accountability
Page purpose	Inform, report, sell, campaign, recruit, entertain or document	Purpose affects what evidence and omissions you should expect
Original evidence	Report, dataset, study, interview, policy text or no traceable evidence	A summary page may be much stronger or weaker than its underlying source
Evidence date	When the evidence was collected or published	A recently updated footer does not make old evidence current
Independent coverage	Who else describes the organisation, evidence or claim	Lateral checking can confirm context, expose disagreement or reveal uncertainty
Assignment relevance	Population, setting, construct, jurisdiction and required source type	A trustworthy page can still be the wrong evidence for the task
Decision	Use, use with limits or exclude, with a one-sentence reason	The evaluation must change how you use the source

The Association of College and Research Libraries Framework treats authority as constructed and contextual. Credentials or format may be relevant, but they do not settle whether a source is appropriate for the current information need. Cornell University Library’s credibility guidance similarly brings together currency, relevance, authority, evidence and purpose.

Identify responsibility without relying on the URL

Start with the person or organisation that takes responsibility for the page. This is not always the same as the site owner or the institution named in the domain.

Check:

• the byline and author biography;
• the responsible department or organisation;
• editorial or review information;
• ownership, funding or sponsorship;
• relevant qualifications or direct experience;
• contact and correction information; and
• whether independent sources confirm the claimed identity or role.

A missing personal byline is not automatically disqualifying. Organisations can be responsible authors, and some reference styles explicitly allow group authorship. The more important question is whether responsibility is identifiable and appropriate for the claim.

A familiar .edu, .gov or .org suffix is also not a quality certificate. It may tell you something about the domain holder, but not who wrote a particular page, how it was reviewed or whether its evidence fits your assignment.

Trace the claim to the original evidence

When a page presents a number, causal claim or research conclusion, follow the evidence chain.

Ask:

• Does the page link to the report, study or dataset?
• Is the link to the original source or another summary?
• Does the original use the same measure, population and time period?
• Does the design support the strength of the wording?
• Have qualifications disappeared between the original and the web page?
• Can you cite the original evidence directly?

This step often changes the decision. A landing page may say that a programme “raised retention”, while its report measured participants’ intention to continue. A news story may state that one factor “causes” an outcome, while the linked study reports an association. A current webpage may rely on a report that is now outdated for a fast-changing topic.

If the original source is available and suitable, cite it rather than the promotional or secondary page for the research finding. Keep the web page only when its wording, public position or communication strategy is itself relevant.

Check currency at the level of the evidence

Website dates can refer to several different events:

• the first publication of the page;
• a superficial page update;
• the release of the linked report;
• the period when data were collected; or
• the date of a policy or legal rule.

Record the date that matters to the claim. “Updated 2026” is not enough if the underlying data were collected in 2018.

Currency is also contextual. Older evidence may remain appropriate for a historical claim, a foundational theory or a stable definition. It may be weak for a current service price, regulation, software feature, public-health recommendation or rapidly changing statistic. Explain what the age changes instead of applying a universal cutoff.

Separate purpose and interest from truth

A source’s purpose influences how you read it. Commercial, advocacy and institutional sources may select evidence that supports their goals. That interest deserves explicit attention, but it does not make every statement false.

Record the relationship precisely:

• The organisation sells the intervention being evaluated.
• The campaign page seeks support for a policy.
• The university page recruits students to the programme it describes.
• The professional body sets standards for its members.
• The author discloses funding from an interested organisation.

Then decide what follows. You may need the underlying evidence, independent comparison, narrower attribution or a different source. Avoid the shortcut “biased, therefore unusable”. A campaign page may be valuable primary evidence of the campaign’s position even when it is unsuitable as neutral proof of the policy’s effects.

Work through a complete fictional example

The following organisations, pages, reports, counts and findings are invented for teaching.

A student is writing about whether first-year belonging programmes improve university retention. They find a 2026 landing page titled “Remote mentoring raises first-year retention by 18%”. It is published by the fictional LearningBridge Institute, which sells the mentoring platform. The page links to a fictional 2022 impact report.

Step 1: record observations

Check	Fictional observation
Intended claim	Remote mentoring improved first-year retention
Responsible party	LearningBridge Institute; no individual research author is named
Purpose and interest	The institute promotes and sells the intervention
Original evidence	A 2022 report surveys 420 voluntary programme participants
Outcome measured	Intention to continue studying, not enrolment records
Comparison	No non-participant group and no before-and-after retention comparison
Currency	Page updated in 2026; underlying report published in 2022
Lateral check	Independent sources confirm the institute exists but do not corroborate the 18% retention claim
Assignment relevance	The assignment concerns belonging; the page reports a retention claim

Step 2: explain what those observations change

The commercial interest calls for scrutiny, but it does not prove the page false. The more decisive problem is the evidence chain. Voluntary participants’ intention to continue is not the same measure as actual retention. Without a suitable comparison, the report also cannot establish that mentoring caused an improvement.

The assignment asks about belonging, which is another construct again. Even a well-supported retention finding would not directly establish students’ sense of belonging.

The lateral search has not independently corroborated the 18% claim. That should be recorded as not independently corroborated, not rewritten as false.

Step 3: make a bounded decision

Proposed use	Decision	Reason
Evidence that remote mentoring increased retention by 18%	Exclude	The report does not measure recorded retention or provide a comparison that supports improvement
Evidence that remote mentoring increased belonging	Exclude	Neither the landing page nor the report measures belonging
Background evidence that mentoring programmes exist	Use with limits	The page documents one provider’s programme, but should not be generalised to the sector
Primary evidence of how the institute promotes its platform	Use	The page directly documents the institute’s own public promotional claim; attribute it to the institute

The same page therefore receives different decisions for different claims. Evaluation is not a permanent label attached to the URL.

Use three decisions instead of a vague credibility score

A numerical score can conceal the reason a page succeeds or fails. Use an explicit decision tied to the intended claim.

Use

Choose use when the page is responsible, traceable, current enough, relevant and suitable for the job you have assigned it. State whether you are citing the page itself or an original source it helped you locate.

Use with limits

Choose use with limits when the page can support a narrower proposition, provide context or serve as primary evidence of an organisation’s position, but cannot carry the broader claim. Write the limit into your notes so it survives drafting.

Exclude

Choose exclude when the page does not meet the assignment’s source requirements, the evidence cannot be traced, the measure or population does not match, the claim exceeds the design, the material is too outdated for the purpose or crucial details remain unverifiable.

Excluding a page from citation does not mean it had no value. It may have supplied vocabulary, led you to an original report or revealed a public position. Record that discovery role separately from evidentiary use.

Avoid common website-evaluation shortcuts

Shortcut	Why it fails	Better move
“It is a .org, so it is reliable”	A suffix does not establish authorship, review or evidence quality	Identify responsibility and trace the claim
“It looks professional”	Design quality is not research quality	Inspect evidence, purpose and independent coverage
“There is no named person, so nobody is responsible”	Organisations may be responsible authors	Identify the group, department or publisher
“It is biased, so everything is false”	Interests shape selection but do not decide every claim	Name the interest and test the evidence
“Several sites repeat the statistic”	Repetition may come from one press release	Trace the statistic to its earliest available source
“The page says updated this year”	The underlying evidence may be much older	Record the evidence date and data period
“I cannot confirm it, so it is false”	Lack of corroboration may leave uncertainty	Record the verification limit and avoid an unsupported claim
“A credible source supports any claim on the topic”	Relevance depends on measure, population and inference	Test the exact evidence–claim fit

Move the decision into your research record

Do not leave the reasoning in browser tabs. Add the evidence card to your source notes with:

• the page title, responsible party, URL and access date;
• the exact claim you considered;
• links or identifiers for the original evidence;
• the evidence date and relevant population or setting;
• the result of the lateral check;
• the use, limited-use or exclude decision; and
• any wording you must preserve when attributing a position.

If the page survives evaluation, the next task may be to write an annotation. Use the guide to evaluating a source in an annotated bibliography to turn the evidence card into a selective summary, criterion-based judgment and bounded plan for use.

A reusable website-evaluation sequence

A good evaluation does not guarantee that a page is true in every respect. It gives you a transparent reason for the precise academic use you make of it—and for the uses you refuse.

If you want guided feedback on how you are selecting and using sources in an argument you have designed, our academic writing development & revision service can help you test evidence use, structure and clarity while keeping the research decisions and authorship yours.