PSLE Science Reality Lab Vol No.441 | “Open Access Article” — Does That Mean the Research Data Are Open Too?

Series ID: PSLE-SCI-REALITY-0441

Wait, What? You Can Read the Paper for Free and Still Not Have the Data

A student finds a scientific article with a bright OPEN ACCESS label. The full paper opens without a subscription. There is no paywall. The student says, “Great. That means the raw research data must be open too.”

Not necessarily.

Open access is mainly an access and reuse condition for the publication itself. UNESCO describes open access as making information freely available online and allowing broad lawful use under appropriate licensing. Open research data are a related but separate object: the measurements, observations, images, code or other evidence produced or used in research may have their own repository, licence, documentation and access conditions.

A paper can therefore be fully open while some or all of its underlying data are restricted, unavailable, available only on request, deposited somewhere else, or open under a separate record. The reverse can also happen: a dataset may be openly available even when an article discussing it is not freely accessible.

The PSLE Science learner’s job is to keep the publication and the evidence objects behind the publication separate.

Quick Answer

No. “Open Access Article” does not automatically mean the underlying research data are open.

Open access tells you that the publication can be accessed freely under stated terms.
Open data tells you that research data are available for access and reuse under stated conditions.
They can travel together, but one label does not prove the other.
To find the data, look for a data-availability statement, repository link, dataset DOI or supplementary files.
If data are restricted, ask why and what evidence remains available for checking the claim.

The quiet rule is: open page ≠ open evidence trail.

The Exact Learner Job This Reality Lab Owns

This page owns one evidence-transfer job: how a Primary 5/6 learner interprets the “Open Access” label on a scientific publication without assuming that the underlying data, code, samples or complete evidence record are also open.

It does not replace the existing Reality Lab owner for Open Data Available. That owner asks what open data allow a reader to check and why open data do not mean a result has already been reproduced. This article stops one step earlier: it asks whether the data are actually open at all when the article is open access.

The Two-Door Model

Imagine a research project with two doors.

Door A: Publication access. Can you read the article?
Door B: Evidence access. Can you inspect the underlying data or other research objects?

Opening Door A does not automatically open Door B.

This simple separation prevents several common mistakes. A learner no longer assumes that a free PDF includes the raw spreadsheet. A learner also stops treating restricted data as proof that the article is false. Instead, the learner asks what is accessible, what is not, why, and how that changes the strength of the check they can perform.

Observed, Claimed and Inferred

Suppose a fictional ecology article has this label:

OPEN ACCESS — Full article available under an open licence

Later, the article says:

Data are available from the research team on reasonable request because exact nest locations are sensitive.

Observed: the article itself is open access.
Observed: the exact location data are not openly downloadable.
Claimed: access to the location data is controlled because of a stated sensitivity concern.
Invalid inference: “Open access was a lie because the raw data are not public.”
Another invalid inference: “The paper is open, so every underlying file must be public.”

Both labels can be accurate because they refer to different objects.

Open Publication, Closed Data: Why Can That Happen?

There are legitimate reasons some research data cannot simply be posted for everyone.

Data may contain personal or private information.
Exact locations of threatened species may need protection.
Indigenous or community data may have justified governance restrictions.
Some records may be covered by legal or contractual limits.
A dataset may need controlled access rather than unrestricted public download.

UNESCO’s Recommendation on Open Science supports making scientific knowledge as open as possible while recognising justified restrictions involving privacy, confidentiality, rights and protection of sensitive information.

For a Primary learner, the scientific habit is not “open good, closed bad”. It is identify the access boundary and ask how it affects what can be checked.

Worked Case 1: The Free Paper With a Hidden Spreadsheet

A fictional open-access paper compares how quickly three materials cool. The paper contains a graph and summary averages but no raw readings. The data-availability statement says the spreadsheet can be requested from the authors.

A student says, “Because the article is open access, I can check every original temperature reading.”

Not yet. The student can inspect the published graph, method and summary values immediately. To inspect every raw reading, the student would need access to the spreadsheet. Open access improves access to the communication object but does not create evidence files that are not attached to it.

The correct question becomes: What can I verify from the open article, and what would require the underlying data?

Worked Case 2: Open Data, Paywalled Article

A public repository contains a freely downloadable dataset of tree heights with a clear licence and documentation. A separate journal article using that dataset is behind a subscription wall.

A learner says, “The data cannot be open because the article costs money.”

Again, the two doors were confused. The dataset can be open even when the paper is not. The learner can examine the measurements and documentation in the repository but may not be able to read the journal’s full interpretation without legitimate access to the article.

Worked Case 3: Supplementary Table Is Not the Whole Dataset

An open-access paper includes a supplementary table containing 40 selected results. The methods section says the full project recorded 2,400 measurements.

A student downloads the supplement and calls it “the open raw data”.

That description may be wrong. A supplementary table could contain processed or selected values rather than the complete raw dataset. The learner should identify:

what the supplementary file actually contains;
whether values are raw, cleaned, transformed or summarised;
whether any records were excluded;
where the complete dataset is described;
which version of the data supports the published analysis.

“Downloadable” is an access fact, not a guarantee about completeness or processing level.

Worked Case 4: Sensitive Wildlife Locations

A study measures nest success for a rare bird. The article is open access. A map shows broad regions, but exact nest coordinates are withheld from public download to reduce risk to the animals.

Two extreme reactions fail.

Reaction A: “No exact coordinates means the study must be fake.”
Reaction B: “The article is open, so we do not need to know anything about the hidden data.”

A better scientific response is proportional. The restriction has a reason. The reader can still examine the reported method, sampling logic, summary results and other available evidence. But some independent checks may be limited unless qualified reviewers or researchers can access the protected coordinates under appropriate conditions.

The Access Map

When reading a scientific article, map the evidence objects rather than asking one vague question such as “Is it open?”

Object	Possible status	Question to ask
Article	Open access / restricted	Can I read the full publication?
Dataset	Open / controlled / unavailable / on request	Can I inspect the observations?
Code	Open / partial / unavailable	Can I inspect how calculations were produced?
Materials or protocol	Shared / described / restricted	Can I understand or repeat the method?
Supplement	Attached / partial / absent	What extra evidence does it actually contain?

The map turns “open” from a vague praise word into a set of testable access statements.

Representation Check: The Open-Access Padlock Is Not a Data-Repository Icon

Websites often use open-padlock symbols, bright badges or licence logos near an article title. These can be useful signals, but the visual design encourages fast inference.

Pause and read the small print. Is the badge attached to the article, the dataset, the journal or the whole website? Does it describe free reading, reuse permissions, public-domain status or something else?

A symbol can compress information, but it can also hide the object boundary if the reader does not ask what the symbol modifies.

Comparison Check: Two Open-Access Papers May Offer Very Different Evidence Access

Paper A and Paper B are both open access.

Paper A links to a documented dataset and analysis code.
Paper B provides a methods section and summary tables but no public dataset because participant privacy prevents open release.

The shared label “open access” does not make their evidence-access arrangements identical. A careful comparison describes what each allows a reader to inspect rather than awarding a single trust score.

Method Check: Open Access Does Not Repair a Weak Investigation

Suppose an article is completely free to read and its data are downloadable. The experiment still uses only one plant, changes two variables at once and makes a conclusion about all plant species.

Openness makes the weakness easier to see. It does not turn the weak design into a strong one.

This is an important distinction: transparency can improve checkability without guaranteeing quality.

Alternative Explanations for Missing Public Data

If data are not openly downloadable, several explanations are possible.

There is a legitimate privacy or sensitivity restriction.
The repository uses controlled access.
The data are available on request.
The data exist but were poorly documented or never deposited.
Only summary data were preserved.
The publication predates stronger data-sharing practices.
The link is broken or the repository record changed.

Do not guess which explanation is true. Read the data-availability statement and repository record.

What Strengthens an Open-Evidence Claim?

A clear data-availability statement.
A direct repository link or persistent identifier.
Documentation explaining variables, units and missing values.
A licence or access condition that is easy to understand.
A version number or date identifying the dataset used in the paper.
Code or calculation steps where relevant.
Transparent explanations for justified restrictions.
Enough metadata for another reader to know what the files represent.

What Weakens the Shortcut “Open Access = Open Data”?

The article is free, but the data statement says “not publicly available”.
The paper links only to summary tables.
The downloadable files are processed outputs rather than raw observations.
The repository contains a different version from the one used in the article.
The open-access licence applies to the article text but not automatically to every third-party object.
No evidence connects the “open” badge to the dataset itself.

How Far Can the Conclusion Travel?

An open-access label can support a narrow conclusion: the publication is openly accessible under stated terms. It cannot travel automatically to:

all underlying data;
all analysis code;
all laboratory records;
all specimens or images;
peer-review quality;
correctness of the conclusion;
independent replication.

Every extra step needs its own evidence.

Tempting but Invalid Reasoning

“Open access means open data.” They are separate access claims about different objects.
“If data are restricted, the study is automatically untrustworthy.” Restrictions can be justified; inspect what review and verification remain possible.
“If the paper is free, it has been checked more carefully.” Access mode does not by itself establish evidence quality.
“Downloadable supplement equals complete raw data.” Identify what the supplement contains.
“Open data means the result has already been independently repeated.” Reusing the same data is not the same as collecting new evidence.

PSLE-Style Transfer Case

A fictional article titled How Shade Affects Playground Surface Temperature is labelled Open Access. The article shows a graph of mean temperature at four sites. Its data statement says:

The complete timestamped measurements are deposited in a public repository. Photographs showing children in the sampling area are not publicly released.

Question 1: Does “Open Access” alone prove the measurements are public?

Answer: No. The separate data statement provides that evidence.

Question 2: Does withholding the photographs mean the temperature data cannot be checked?

Answer: Not necessarily. The temperature dataset is publicly deposited, while the photographs have a separate privacy boundary. We should evaluate whether the available measurements and metadata are enough for the scientific claim.

Question 3: Does public access prove the shade caused the temperature difference?

Answer: No. The method and alternative variables still have to be evaluated.

Explained Practice: Name the Object After the Word “Open”

Repair each vague statement by naming the object.

“The research is open.” → Is the article open, the data open, the code open, or all three?
“Anyone can check it.” → Which part can anyone inspect?
“The files are available.” → Which files, which version and which processing level?
“The data are restricted.” → Which data and what is the stated reason?
“The study is transparent.” → Transparent about method, data, code, exclusions, uncertainty, or some combination?

Naming the object turns a slogan into a scientific description.

Delayed Independent Return

On another day, draw two doors labelled Article and Data. Create four fictional studies: both doors open; article open/data controlled; article restricted/data open; both restricted. For each case, write what a learner can inspect and what remains unavailable.

Then add a strong scientific claim. Ask whether the access pattern itself proves the claim. It never does. Access changes what can be checked; evidence and reasoning still determine what the claim deserves.

Useful eduKateSengkang Routes

Parent and Tutor Teaching Guide

Prepare four cards: ARTICLE, DATA, CODE and IMAGES. Put an “open”, “controlled” or “not supplied” sticker on each card. Ask the learner to describe the evidence-access pattern without using the vague sentence “the study is open”.

Then give an open-access article with no public data and a restricted article with an open dataset. Ask which one provides more access to measurements. The answer should depend on the object being discussed rather than the article badge.

Finally, add a legitimate restriction such as sensitive wildlife coordinates. Ask what checks are still possible and what checks become harder. This prevents learners from turning openness into a moral score instead of an evidence property.

Authoritative Sources

The Quiet Return

Open access is genuinely useful. It lets more people read scientific work, examine arguments and follow references without a payment barrier.

Its value becomes clearer, not weaker, when we keep its boundary precise.

Ask which object is open. Then trace the evidence behind it. Science becomes easier to evaluate when access labels are treated as descriptions of access rather than shortcuts to truth.

Small Group Tutorials