How to Avoid Counting the Same PSLE Science Evidence Twice When It Appears in More Than One Representation

Wait, What? A question can show the same result three times and still give you only one underlying piece of evidence.

A sentence may state that Set-up A reached 28 units. A table may contain the value 28 for Set-up A. A graph made from that same table may place a point at 28. Those three representations can help you understand the result, but they do not automatically become three independent experiments. If you count each display as fresh support, your conclusion becomes stronger on paper than the evidence really is.

This guide teaches a powerful PSLE Science habit: count evidence by where it came from, not by how many times it is displayed.

Quick Answer

Whenever the same scientific information appears in words, a diagram, a table or a graph, ask: “Is this a new observation or measurement, or is it another representation of an earlier one?” If the graph was drawn from the table, the graph and table usually share the same underlying data. If the sentence simply describes the graph, that sentence may also be a restatement rather than a new observation.

ONE MEASUREMENT
     ↓
  TABLE
     ↓
  GRAPH
     ↓
WRITTEN DESCRIPTION

four displays ≠ four independent measurements

The Exact PSLE Science Learning Job This Guide Owns

This page owns one job: how a Primary 5/6 learner tracks evidence provenance across several representations and avoids double-counting the same underlying result. It does not replace the existing guides on how to read tables, graphs or diagrams. Those teach representation. This guide teaches evidence independence: whether two displayed facts actually come from two separate observations, measurements, trials or sources.

Why Multiple Representations Are Useful—and Why They Can Mislead

Science often represents the same relationship in different forms because each form makes something easier to see. A table preserves exact values. A graph can make a trend visible. A diagram can show arrangement. Words can state conditions or explain meaning. The problem begins when a learner mistakes different views for different evidence sources.

Think of a photograph shown on a phone, a tablet and a printed page. You now have three copies, but not three separate events. Scientific data can behave the same way.

That does not mean repeated representation is useless. It may let you cross-check a transcription, find a trend, identify a mislabeled axis or connect a number to a physical set-up. The key is to separate two questions:

How many representations do I have?
How many independent observations or measurements produced them?

The PSLE Science Reasoning Chain

READ GIVEN INFORMATION
→ IDENTIFY EACH SCIENTIFIC CLAIM OR VALUE
→ TRACE WHERE EACH ONE CAME FROM
→ DISTINGUISH ORIGINAL EVIDENCE FROM A RE-DISPLAY OR SUMMARY
→ SELECT THE RELEVANT CONCEPT
→ EXPLAIN THE MECHANISM
→ CONNECT THE MECHANISM TO THE ACTUAL EVIDENCE
→ STATE THE OUTCOME
→ CHECK THAT SUPPORT HAS NOT BEEN DOUBLE-COUNTED

This is especially important when a question asks which conclusion is best supported, whether a claim is justified, or what evidence shows a relationship.

Worked Example 1: Table and Graph From the Same Data

Imagine an original practice investigation. Four test conditions produce four measurements: 12, 17, 21 and 24 units. The question first gives the values in a table, then shows a graph constructed from those same four values.

What evidence do you have? You have four underlying measured results, represented twice. You do not have eight measured results.

Displayed item	New underlying evidence?	Why?
Table: 12 units	Yes, if this is the original result record	It represents one measured result.
Graph point at 12 units	No, if plotted from that table	Same result, new representation.
Sentence: “The value was 12 units”	Usually no, if it merely describes the same data	Restatement of the same result.
Second independent trial gives 13 units	Yes	A new trial produced a new result.

The graph can still reveal a pattern that is hard to see in the raw table. But the pattern is derived from the same measurements. It is not an independent replication of them.

Worked Example 2: A Diagram Can Add New Evidence

Now suppose a question contains a results table and a separate diagram of the apparatus. The diagram shows that one thermometer is placed much closer to a heat source than another. Is the diagram merely a duplicate of the table?

No. The diagram may provide a different kind of given information: spatial arrangement or experimental condition. The table provides measured results. They can be combined because they are not simply copies of each other.

This is why you should not use a crude rule such as “table plus graph equals duplicate; diagram equals new.” Instead, trace the scientific job of each item.

Does it report a new observation?
Does it report a new measurement?
Does it establish a condition?
Does it show an arrangement?
Does it simply re-display information already given elsewhere?

Evidence Source vs Evidence Representation

Use these two labels:

Term	Meaning in this guide
Evidence source	The actual observation, measurement, trial, specimen comparison or given fact that provides information.
Evidence representation	The form in which that source is shown: words, table, graph, diagram, symbol or calculated summary.

One source can have several representations. One representation can also contain several sources—for example, a table with results from six genuinely separate trials. So always inspect provenance rather than counting boxes, rows or pictures.

A Useful Test: Could One Display Disappear Without Losing the Underlying Result?

Suppose the table and graph show exactly the same values. If the graph disappeared, would the original measured values still exist in the table? Yes. That is a clue that the graph is a transformation of the same data rather than a new experiment.

Now suppose the diagram supplies a condition that the table never mentions. If the diagram disappeared, would you lose information needed to interpret the results? Yes. Then the diagram is contributing a distinct piece of given information.

When a Calculated Value Is Not Independent Evidence

A question may give two direct measurements and a calculated difference between them. The difference is useful, but it comes from those two measurements. If the measured values are 30 units and 24 units, the difference of 6 units is not a third independent observation.

The calculation can answer a different scientific question—such as how much change occurred—but its evidence provenance still traces back to the original measurements.

Failure Signatures

You say a claim is “strongly proven” because the same number appears in a table and graph.
You count a calculated difference as a separate trial.
You treat a sentence summarising the graph as new evidence.
You ignore genuinely new information in a diagram because you assume all representations are duplicates.
You count five graph points as five trials even though they are five time measurements from one trial.
You cannot explain where a number originally came from.
You use representation count instead of evidence quality to decide which claim is better supported.

Earliest Weak-Link Diagnosis

What went wrong?	Earliest weak link	Repair
Graph + table counted twice	Source tracking	Draw an arrow from graph back to the table/data source.
Calculated value counted as new measurement	Direct vs derived evidence	Label values M for measured and C for calculated.
Diagram’s condition ignored	Representation role	Ask what new information would be lost if the diagram were removed.
Five time points called five trials	Trial vs measurement	Count complete experimental runs separately from readings within one run.
Conclusion stronger than evidence	Evidence independence	Count distinct sources and check method quality before claim strength.

The SOURCE Protocol

This is a practice routine, not an official examination format.

S — Spot every claim, value and observation being used.
O — Origin: where did each one come from?
U — Underlying result: are two displays actually the same result?
R — Role: measurement, condition, observation, calculation, representation or inference?
C — Combine only genuinely complementary evidence.
E — Evaluate the conclusion using evidence quality, not display count.

Original Practice 1: Three Displays, One Dataset

A student measures a quantity every five minutes during one investigation. The results are printed in a table. A line graph is drawn from the table. A paragraph then states that the quantity increased over the first ten minutes and later levelled off.

Ask yourself:

How many experimental runs are described?
How many measurement times occur?
Does the graph add new measured values?
Does the paragraph add new measured values?
What new job does the graph perform even if it is not independent evidence?

A strong answer recognises that multiple time measurements can belong to one run, that a graph can reveal the pattern more clearly, and that a written summary may be an interpretation of the same dataset rather than additional replication.

Original Practice 2: Two Truly Different Evidence Sources

Suppose a question gives a diagram showing the arrangement of two set-ups and a table showing their final measurements. The diagram establishes that only one stated condition differs. The table shows that the measured outcome also differs.

Here the diagram and table play complementary roles. The diagram supports what was different in the method; the table provides the outcome evidence. They should be connected, not collapsed into one source and not counted as two independent repeats of the same result.

MCQ Reasoning: Beware “More Mentions Means More Support”

In an MCQ, an option may feel well supported because several parts of the page seem to agree with it. Before accepting that feeling, trace whether those parts are independent. If the table, graph and sentence all come from the same measurement, a flaw in that measurement would affect all three displays.

Scientific confidence should rise because evidence is relevant, well produced, appropriately repeated and consistent—not merely because a single result has been reformatted several times.

Open-Ended Answers: Use the Decisive Evidence Once

When writing an explanation, you usually do not need to say, “The table shows 28 units, the graph also shows 28 units and the paragraph says 28 units.” That is repetition. State the decisive result once, then use the relevant scientific concept and mechanism to explain it.

This produces a cleaner chain:

relevant condition
→ decisive evidence
→ scientific concept
→ causal mechanism
→ outcome / conclusion

Misconception Repair: “If It Is Shown Again, It Confirms It”

Repetition of a representation is not the same as repetition of an experiment. A graph drawn from a table can confirm that the plotting was done consistently with the table, but it cannot independently confirm that the original measurement was accurate. The source chain matters.

Repair sentence: “A new format can clarify old evidence; only a new valid source adds independent evidence.”

Retrieval and Practice Sequence

Map one dataset: take a table and graph and draw arrows back to the same measurements.
Add one independent repeat: show how a second trial creates genuinely new evidence.
Mix roles: include a diagram that contributes a new condition so the learner cannot simply label every second representation “duplicate.”
Remove a representation: ask what scientific information is lost.
Delay: return after several days with a new question using different representations.

Unfamiliar Transfer Test

A question gives: (1) a photograph of a set-up, (2) a table of measured values, (3) a bar chart made from the table, and (4) a second independent trial whose values are listed in a new row. Which parts add independent evidence about repeatability?

The second independent trial does. The photograph may add method information. The chart helps representation. But the chart does not create another trial from the first table.

Delayed Independent Return Test

Three to seven days later, explain from memory the difference between an evidence source and an evidence representation. Then invent one example in which a diagram is only a re-display and one example in which a diagram contributes genuinely new condition information. If you cannot create both, the boundary is not yet stable.

Answer and Checking Receipts

I can trace each value to its original observation or measurement.
I can tell whether a graph was constructed from a table already shown.
I can distinguish direct measurements from calculated summaries.
I can distinguish repeated measurements from repeated representations.
I can identify when a diagram contributes a new condition rather than restating data.
I can use decisive evidence without listing the same fact several times.
I can keep conclusion strength matched to genuinely independent support.

Common Traps

Trap: “The graph confirms the table.” Check: was the graph made from that table?
Trap: “Three displays mean three experiments.” Check: count actual trials and observations.
Trap: “Calculated means measured.” Check: trace the calculation to its source values.
Trap: “All representations are duplicates.” Check: some may add conditions, identities or arrangements.
Trap: “More mentions mean stronger evidence.” Check: evidence quality and independence matter more than repetition of display.

Parent and Tutor Teaching Guide

Give the learner one small dataset and display it three ways. Ask the child to draw arrows back to the original measurements. Then introduce a genuinely independent second trial and ask what changed in the evidence structure.

When a learner says, “There is lots of evidence,” ask, “How many separate observations or trials created it?” This question develops scientific caution without teaching distrust. The goal is not to make the child suspicious of graphs. The goal is to understand what graphs are evidence of and where their data came from.

EEF’s primary science guidance emphasises teaching pupils to work scientifically alongside building coherent science knowledge. Evidence-source tracking is a practical way to make that relationship visible: pupils must understand both the science and how the evidence was produced.

Useful Internal Routes

Official Frame and Authoritative References

The 2026 PSLE Standard Science syllabus assesses the 2023 Primary Science syllabus and includes interpretation and analysis of information, evaluation of observations and methods, and communication of explanations and reasoning. This guide uses those official aims as its boundary. It does not claim that “evidence independence” is a named PSLE marking rubric or that students must use the SOURCE acronym in an examination.

Quiet Return

Science becomes clearer when every claim can answer a simple question: “Where did this evidence come from?” A table, graph and sentence can all be valuable, but their value is not the same as independent replication. Trace the source, protect the meaning, then let the real evidence carry exactly the weight it deserves.

Small Group Tutorials