PSLE Science Reality Lab Vol No.009 | “Every Trial Worked” — Were the Failed Trials Left Out?

Series ID: PSLE-SCI-REALITY-0009

Wait, What? “Five Successful Trials” Does Not Tell You Whether Five Trials Were Attempted

A short video montage shows five attempts. In every clip, the result works beautifully. The caption says:

“It worked every time.”

There is one missing number that can completely change what that sentence means: how many attempts were made in total?

If five trials were attempted and five succeeded, the observed success record is five out of five. If twenty trials were attempted and only the five successful ones were shown, the visible montage is still made of real successes—but it no longer represents the full record.

This is a powerful transfer of PSLE Science inquiry into everyday life. The 2026 PSLE Science assessment objectives include evaluating observations, information and methods. The 2023 Primary Science syllabus asks learners to exercise healthy scepticism, question assumptions and uncertainty, assess the quality of information and understand how Science is communicated through different media. A missing result is therefore not just a media problem. It is an evidence problem.

Quick Answer

When a claim reports only successful attempts, ask:

How many total trials were attempted?
What counted as a success before the trials began?
Were unsuccessful, ambiguous or partial results recorded?
Were the conditions the same across attempts?
Were any trials excluded, and why?
Does the complete record support the same conclusion as the selected successes?

Missing information is not proof of dishonesty. But it limits what a careful learner can conclude.

The Owned Learner Job

This Reality Lab owns one transfer job: how to evaluate a real-world “all our trials worked” or “every example succeeded” claim when the communication may show selected successes without the full set of attempts.

It does not replace the existing eduKateSengkang guides on repeated trials, anomalous results, evidence limits, internal consistency or fair testing. Those remain the canonical owners of the individual reasoning skills. This page combines them around a communication pattern learners often meet in videos, product demonstrations and persuasive claims.

Reality Lab Case: The Paper-Bridge Challenge

Imagine an original video series testing a folded paper bridge. The presenter places ten identical coins on the bridge in five separate clips. In all five clips, the bridge holds.

The caption says, “Our design held ten coins in every trial.”

If those were the only five trials attempted, the statement accurately describes the record. But suppose the presenter actually performed twelve trials. In seven earlier trials, the bridge collapsed because folds were slightly uneven. Only the five successes were included in the montage.

Now several claims must be separated:

True: the design held ten coins in each of the five displayed clips.
Not established: the design held ten coins in every attempt.
Possible but needs analysis: the design can be reliable if folded within certain tolerances.
Too broad: this paper-bridge design always holds ten coins.

The selected clips are not fake. The problem is that the selected set may not represent the full attempt set.

The Complete-Record Question

In a PSLE Science investigation, learners are taught to record observations and results. Reality Lab adds one question whenever the evidence arrives already packaged for an audience:

Am I seeing the result record—or only a selection from the result record?

This question matters because selection can change frequencies, averages, apparent consistency and the range of outcomes.

Success Must Be Defined Before It Is Counted

Suppose ten seeds are treated. A post says eight “grew well”. What does that mean?

Did they germinate?
Did they reach a certain height?
Did they grow taller than untreated seeds?
Did they survive for a set number of days?
Did the presenter decide what “grew well” meant only after seeing the results?

A success criterion should match the scientific question and be clear enough to apply consistently. If the definition changes after the results are known, the count becomes difficult to interpret.

Missing Results and Anomalous Results Are Not the Same Thing

An anomalous result is a recorded result that does not fit the main pattern. A missing result is one we do not have. They require different responses.

For an anomaly, investigate possible measurement error, procedure differences, natural variation or a real boundary condition. Do not delete it merely because it is inconvenient.
For a missing result, first ask why it is missing. Was the instrument unreadable? Was the trial interrupted? Was it excluded by a rule decided beforehand? Was it simply not reported?

The absence of information creates uncertainty. It does not tell you automatically what the missing result would have been.

Worked Case 2: The “Perfect” Cooling Test

A student compares two cup covers and wants to know which keeps water warm longer. A social post shows four trials where Cover A gives a higher final temperature than Cover B. The post says, “Cover A wins every time.”

Ask for the full record. Suppose there were six trials:

Trial 1: A higher
Trial 2: A higher
Trial 3: readings missing because the thermometer slipped
Trial 4: B slightly higher
Trial 5: A higher
Trial 6: A higher

The four displayed successes are real. But “wins every time” is inaccurate because one completed trial favoured B, and another trial did not provide a usable comparison.

A more defensible statement would be: “Cover A produced a higher final temperature in four of the five trials with usable paired readings.” Then we should still inspect whether starting temperatures, water amounts, cup types, timing and thermometer placement were comparable.

Why “We Repeated It” Is Not Enough

Repeating trials can improve evidence, but only if we know what happened across the repetitions. A montage of successful repeats can create the appearance of reliability while hiding the actual variation.

Scientific repeatability asks how consistently an effect appears under stated conditions. Consistency cannot be judged from successes alone if failures were also part of the same test series.

A Simple Evidence Ledger

For any “every trial worked” claim, rebuild a small ledger:

Total attempts: ___
Usable results: ___
Successes: ___
Non-successes: ___
Missing/invalid trials: ___
Reason for exclusion: ___
Same conditions across trials? yes / no / unknown

If several boxes are unknown, say so. Unknown is a valid scientific state. It is better than inventing completeness.

Selective Reporting Can Happen Without Deliberate Deception

People naturally like clean stories. A teacher may show the clearest demonstration because lesson time is short. A student may record the neatest photograph. A video creator may remove failed attempts because they are boring. A company may highlight the strongest test because it is persuasive.

The scientific consequence is similar even when the intention differs: the audience may see a cleaner pattern than the complete evidence contains.

Therefore, do not begin with an accusation. Begin with a methodological question: what was the complete set of attempts?

What Would Strengthen the Claim?

The total number of attempts is stated.
All results are reported or exclusions are explained.
The success criterion was defined before testing.
Conditions were kept comparable across trials.
Missing readings are distinguished from failures.
Repeated results show a stable pattern rather than only selected highlights.
The conclusion matches the actual success frequency instead of using absolute words such as “always” when exceptions occurred.

What Would Weaken It?

Only successful clips are available and the total number of attempts is unknown.
Unsuccessful trials were discarded because they did not fit the expected result.
“Success” was redefined after seeing the data.
Conditions changed between trials but the results were combined as though they were comparable.
Ambiguous results were counted as successes without a prior rule.
The communication uses “every”, “always” or “100%” even though the complete record contains exceptions.

A PSLE-Style Transfer Case

A student tests whether a paper helicopter design stays in the air longer than another design. The student reports only the three longest flight times from each design and concludes that Design P is better.

What is the earliest weak link?

Reasoning: selecting only the longest times may not represent the complete performance of either design. The student should compare results using a method decided before testing and include the full set of valid trials, while explaining any exclusions. If Design P sometimes flies very long but often falls quickly, the conclusion may differ from one based only on best attempts.

Notice how the same reasoning travels from a PSLE-style investigation to a polished online montage.

Delayed Independent Return

Two days from now, take any harmless “works every time” claim and write four numbers or labels:

total attempts,
reported attempts,
success criterion,
excluded or missing results.

If the source does not provide them, do not guess. Write “unknown” and state exactly how that limits the conclusion.

Useful eduKateSengkang Routes

Parent and Tutor Teaching Guide

When a child sees a string of successful demonstrations, ask one neutral question: “How many attempts do you think they made altogether?” Then ask, “Do we know, or are we guessing?”

This helps the learner notice the difference between visible success and complete evidence without turning the lesson into suspicion. If the full record is available, calculate the outcome from all valid trials. If it is not, practise writing a bounded conclusion: “The displayed trials succeeded, but the success rate across all attempts cannot be determined from the information provided.”

For advanced practice, give the learner the same ten-trial record in three versions: all results, only successes, and only failures. Ask what each selection would make an audience believe. Then restore the complete record.

Authoritative Sources

The Quiet Rule to Keep

A row of successes tells you that those successes happened. It does not tell you whether they were the whole experiment.

Before accepting “every trial worked”, find the denominator of experience: every trial that was actually attempted. Science becomes trustworthy not by hiding messy results, but by letting the complete record decide how strong the claim deserves to be.

Small Group Tutorials