How to Tell Whether Two PSLE Science Observations Are Independent Checks or Two Effects of the Same Event

Wait, What? Two observations do not always give you twice as much evidence.

A PSLE Science question may show two pieces of information that look different: a colour changes and a pointer moves; a diagram changes and a table records a value; two measurements are taken from the same specimen; two visible effects appear after one event. It is tempting to count every observation as a separate confirmation of your explanation.

Sometimes that is reasonable. Sometimes both observations are linked to the same underlying event, same specimen state, same measurement chain or same source result. If one observation would automatically change when the other changes, they may not be fully independent checks.

Quick Answer

Before saying “there are two pieces of evidence”, ask: Where did each observation come from? If they arise from different measurements, specimens, trials or independently observed features, they may add separate support. If both are consequences of the same event or one is calculated, copied or derived from the other, treat them as linked evidence and avoid counting them twice.

COUNT EVIDENCE BY ITS SOURCE, NOT BY HOW MANY SENTENCES, NUMBERS OR EFFECTS YOU CAN POINT TO.

The PSLE Science Learning Job This Guide Owns

This guide owns one precise learner job: judging whether apparently separate observations provide independent scientific support or are linked consequences of the same underlying source. It does not teach advanced statistics. It does not replace existing guides on repeated trials, duplicate representations or direct and indirect evidence. It teaches a simpler Primary-level question: “Are these really separate checks?”

The 2026 PSLE Science assessment frame includes interpreting and analysing information, evaluating observations and information, and communicating explanations and reasoning. Evidence quality therefore depends not only on how much information is shown, but on how that information was produced.

Start With an Evidence Family Tree

For each observation, mentally trace backwards:

What object or specimen produced it?
Which trial or time state did it come from?
Which measuring method or observation rule produced it?
Was it measured directly, calculated from another value, or displayed in another representation?
Would this observation still exist as a separate check if the other observation were removed?

The fifth question is especially useful. If Observation B exists only because Observation A was transformed into a graph, percentage or label, B is not a new source of evidence.

Worked Example 1 — One Result, Two Representations

A table records that Set-up P produced 18 units. A graph then plots the same value at 18. The learner writes, “The table and graph both prove P produced 18 units, so there are two pieces of evidence.”

The table and graph are two representations of one underlying measurement. They help the reader understand the evidence, but they do not create a second independent observation. The evidence source is still one recorded result.

Worked Example 2 — Two Effects From One Event

In a fictional system, one event causes both an indicator to change colour and a connected pointer to move because both respond to the same change in the system. The learner treats colour change and pointer movement as two completely independent confirmations.

They are two observations, but their dependence matters. If both are downstream effects of one common event, they do not necessarily provide the same strength of support as two separate tests using independent methods. You may still use both observations if relevant; just do not exaggerate what their number alone means.

Worked Example 3 — Two Independent Measurements

Suppose a question gives two separate measurements from two independently repeated trials carried out under the same condition. Both show the same broad pattern. Here, the second trial contributes new evidence because the procedure was restarted and a new result was generated.

This does not make the conclusion certain. Repeated trials can still share a design flaw. But the second trial is more independent than rereading the same display twice.

Worked Example 4 — Same Specimen, Two Readings

A learner measures the same specimen three times without changing its state. The readings are close. This may tell us something about reading consistency, but the three measurements are tightly linked to one specimen and one physical state. They are not equivalent to testing three separate specimens or repeating the entire investigation three times.

Ask what actually restarted. If only the observer looked again, the evidence structure is different from a new trial.

Worked Example 5 — One Observation Calculated Into Another

A mass difference is calculated by subtracting a final mass from a starting mass. The learner then writes that the final mass and the calculated mass loss are two independent measurements supporting the same claim.

The mass loss depends mathematically on the starting and final measurements. It may be the most useful quantity for the question, but it is derived evidence, not a wholly separate observation. Trace the calculation back to its source readings.

Independent, Linked or Duplicated?

Evidence pair	Relationship	How to treat it
Same measurement in table and graph	Duplicated representation	One underlying evidence source
Two effects caused by one event	Linked consequences	Use both if relevant, but preserve dependence
Two repeated full trials	More independent	Second trial adds new evidence about consistency
Measured value and calculation derived from it	Derived relationship	Do not count calculation as a new measurement
Two different methods checking the same claim	Potentially complementary	Ask whether they share the same hidden limitation

Evidence Independence Is Not All-or-Nothing

At Primary level, you do not need formal probability or statistical dependence. It is enough to recognise degrees of connection. Two results can be more separate or more linked depending on how they were produced.

A useful question is: What failure could make both observations wrong at the same time? If one faulty instrument, one incorrect label or one method mistake would affect both, the observations share a weakness. If the evidence comes from genuinely different routes, one failure may not explain both.

Failure Signatures

“The table and graph both show it, so that is two confirmations.”
The same measurement is counted once as a reading and again as a calculated value.
Several observations from one specimen are described as several specimens.
Two consequences of one event are treated as two unrelated causes.
A repeated display is mistaken for a repeated trial.
The learner counts evidence items without tracing where they came from.

Earliest Weak-Link Diagnosis

Ask the learner to draw arrows backwards from each observation to its source. If two arrows meet at the same measurement, same event or same specimen state, label that shared source. Then ask whether another genuinely new observation exists.

List the observations.
Name the source of each.
Mark which are measured and which are derived.
Mark which share a trial, specimen, method or event.
State what separate support remains.
Build the conclusion at a strength that matches that evidence structure.

Misconception Repair

“More observations always mean stronger evidence.” Not automatically. Ten readings copied from one sensor display are not ten independent experiments.

“If observations look different, their sources must be different.” A colour change and a graph point may both come from one underlying event.

“Linked evidence is useless.” No. Linked observations can be highly informative. The problem is only over-counting them or pretending they are more independent than they are.

“Two independent results prove the explanation.” Independence strengthens the evidence structure, but scientific claims still depend on fair design, relevant concepts, method limits and the range tested.

The Evidence-Provenance Protocol

OBSERVE → NAME THE EVIDENCE ITEM → TRACE ITS SOURCE → CHECK WHETHER ANOTHER ITEM SHARES THAT SOURCE → SEPARATE MEASURED FROM DERIVED → IDENTIFY GENUINELY NEW SUPPORT → APPLY THE SCIENTIFIC CONCEPT → STATE A BOUNDED CONCLUSION.

Practice Sequence

Source tracing: Sort ten evidence items into measured, derived, duplicated or independently repeated.
Same-event test: Given two observations, decide whether one common event could generate both.
Remove-one test: Ask whether Evidence B still exists if Evidence A is removed.
Failure-sharing test: Identify one error that could affect both observations.
Transfer: Repeat using diagrams, tables, repeated trials and qualitative observations across different Science contexts.

Unfamiliar Transfer Challenge

A fictional investigation produces a reading of 30 units. A bar chart displays 30. A sentence says “30 units were recorded.” A second full trial gives 31 units.

How many underlying results are there? Two: 30 from the first trial and 31 from the second.
Are the bar chart and sentence new measurements? No; they repeat the first result.
Does 31 add new evidence? Yes, because it comes from a new trial.
Does agreement between 30 and 31 prove the whole method is valid? No; both trials could share the same design weakness.

Delayed Independent Return Test

After several days, give the learner a fresh question containing a table, a graph, a calculated value and two trials. The learner passes if they independently identify the number of underlying evidence sources rather than counting every displayed item.

Answer-Checking Receipt

I can trace every observation to its source.
I know which values were directly measured and which were derived.
I know whether two observations share one event, specimen or trial.
I do not count the same measurement again because it appears in another representation.
I know whether a repeated reading is a new trial or only another look at the same state.
I can say what failure could affect several evidence items together.
My conclusion matches the true number and quality of evidence sources.

Parent and Tutor Teaching Guide

Use the phrase “Where did this evidence come from?” repeatedly. Give the child three cards that all show the same measurement in different forms and one card from a second trial. Ask them to group the cards by source. This makes evidence provenance concrete without requiring advanced terminology.

Then make the task harder. Give two different observations that both arise from one event. Ask whether they are useless, duplicated or linked. The best answer is usually “linked”: they are distinct observations but not completely separate sources.

Finally, ask what later independent observation would strengthen the claim. This turns evidence evaluation into a forward scientific question rather than a counting exercise.

Useful Internal Routes

Authoritative and Research References

Quiet Return

Strong scientific reasoning does not ask only, “How many observations do I have?” It asks, “How many genuinely new routes to the claim do these observations provide?” Once you trace evidence back to its source, duplicated displays, linked effects and independent checks become easier to tell apart—and your conclusion becomes more honest.

Small Group Tutorials