PSLE-SCI-REALITY-0109
Wait, What? “95% Recovery” Can Mean Almost the Opposite of “95% Removed”
A laboratory summary shows a confident green number:
Spike recovery: 95%
A social-media graphic then announces, “The process removed 95% of the substance.”
That interpretation can be completely wrong.
In many laboratory quality-control settings, a spiked sample is a sample to which a known amount of the target substance has deliberately been added. The laboratory then checks how much of that known addition its method can recover in the measurement. A 95% spike recovery usually means the method recovered about 95% of the known amount that was deliberately added under those test conditions. It does not automatically mean that 95% of the substance originally present in the sample was removed, destroyed, filtered or treated.
The Reality Lab habit is to ask: what exactly was “recovered”, from what baseline, and for what scientific purpose?
Quick Answer
- Find out whether the percentage refers to a deliberately added known amount.
- Separate the original sample result from the spiked-sample result.
- Ask how much was added and how much extra the method measured after the addition.
- Interpret spike recovery as evidence about method performance under those conditions, not as treatment efficiency unless a separate experiment actually tested treatment removal.
- Do not turn one QC percentage into a health, safety, environmental or product-performance claim that it was never designed to support.
The Exact Learner Job This Page Owns
This page owns one real-world evidence-transfer problem: a laboratory quality-control percentage is copied into an infographic, advertisement or explanation and mistaken for a percentage of the original substance that was removed or recovered by a real-world process.
Reality Lab does not become an analytical-chemistry textbook. It leaves specialist chemistry, instrument mechanisms and laboratory method design with their canonical owners. The Primary 5/6 job is narrower and highly transferable: identify the denominator, reconstruct the comparison and keep the conclusion attached to the experiment that actually produced the percentage.
- Reality Lab Vol No.072: “Removes 99%” — 99% of What Was There at the Start, and What Is Left?
- Reality Lab Vol No.094: “Detected” — Does That Tell You How Much Is There?
- Reality Lab Vol No.107: “QA/QC Passed” — Does That Mean Every Result Is Correct?
- How to Tell a PSLE Science Method Limitation From a Mistake in the Investigation
Original Reality Lab Case: The Known Addition
This teaching case uses fictional Substance Q and constructed data. It is not a copied laboratory procedure.
A laboratory first analyses an ordinary sample and obtains a result of 30 units of Substance Q. A second, equal portion of that same sample receives a known addition of 20 units of Substance Q before analysis. The spiked portion is then measured as 49 units.
| Evidence object | Result |
|---|---|
| Original sample | 30 units |
| Known amount added to second portion | 20 units |
| Spiked portion measured result | 49 units |
The extra signal measured after the known addition is 49 − 30 = 19 units. The laboratory added 20 known units and measured an increase of 19 units. In this simplified teaching case, the spike recovery is therefore 19 ÷ 20 × 100% = 95%.
What was recovered? The known added amount, as seen through the measurement method. Nothing in this calculation says that 95% of the original 30 units was removed from the sample.
Observed, Calculated, Claimed and Inferred
| Layer | Statement |
|---|---|
| Observed/measured | The original portion measured 30 units; the spiked portion measured 49 units. |
| Known experimental action | 20 units were deliberately added to the second portion. |
| Calculated QC result | The method recovered 19 of the 20 added units in this simplified case: 95% spike recovery. |
| Unsupported claim | “The process removed 95% of the original Substance Q.” |
| Why unsupported | No treatment-removal comparison was performed by this spike calculation. |
Why Scientists Add Something on Purpose
At first, deliberately adding the target can sound strange. Why make a sample “less natural” if the goal is to measure it?
Because the added amount is known. That known addition gives the scientist a challenge with an answer. If 20 units are added, the method should be able to detect approximately that extra amount if the measurement chain is working appropriately for that sample and method.
EPA quality-assurance guidance describes spiked samples as samples containing a known added amount of the target. Percent recovery of the added material is used as evidence about analytical accuracy. The important Primary-level idea is that the known addition is a test of the measuring process.
The Denominator Check: 95% of What?
Percentages are dangerous when the denominator disappears. “95%” sounds complete, but a scientific reader asks what quantity the percentage is relative to.
| Percentage phrase | Possible denominator |
|---|---|
| 95% spike recovery | The known amount deliberately added for QC |
| 95% removal efficiency | The amount present before a treatment step |
| 95% of samples detected | The number of samples tested |
| 95% recycled content | A defined mass or material basis of the product |
The same number can describe completely different scientific objects. Reading the denominator is part of reading the science.
The Baseline Check: What Is Being Compared With What?
Spike recovery compares the extra amount measured after a known addition with the amount that was deliberately added. Treatment efficiency usually compares a quantity before treatment with the quantity after treatment. These are different experimental designs.
If a public graphic takes a spike-recovery result and labels it “removal”, it has silently changed the baseline. That is an evidence-transfer error.
The Representation Check: Why the Word “Recovery” Is Easy to Misread
In ordinary English, recovery can mean getting something back: recovering lost material, recovering from illness, recovering a recyclable resource. In laboratory quality control, “recovery” can have a narrower technical meaning: how much of a known added amount is accounted for by the analytical method.
A strong science infographic should therefore label the object clearly. “Spike recovery = 95%” is far better than a large unexplained “95% recovered” badge.
The Method Check: Why Can Recovery Be Less Than 100%?
Real measurement chains are not perfect. Some of the added material may not be extracted completely, may interact with the sample material, may be affected during preparation, or may produce a measurement response that differs from a clean standard. Instrument and sampling variability also contribute.
The scientific lesson is not that every recovery below 100% is a failure. Methods establish their own performance expectations for particular sample types and purposes. Reality Lab deliberately avoids inventing a universal “good recovery” range.
Can Spike Recovery Be More Than 100%?
It can, because measurements vary and other effects can make the apparent increase larger than the known addition. A value above 100% does not mean the laboratory created matter from nothing. It tells us to interpret the measurement process, background correction, interference and variability carefully.
This is another useful protection against magical thinking: a percentage in a scientific report is a property of a defined calculation, not a direct physical object floating in the sample.
Alternative Explanation: The Sample Matrix Changes the Measurement
A target substance may be easy to measure in clean water but harder to measure in muddy water, juice, soil extract or another complex material. Other substances in the sample can alter preparation or instrument response. Scientists call the surrounding material the sample matrix.
For a Primary learner, the exact term is less important than the idea: the same target can be harder to measure when it is surrounded by different material. A spike added to the real sample can help reveal this.
What Evidence Would Strengthen a Spike-Recovery Interpretation?
- The report states how much of the target was deliberately added.
- The original unspiked sample result is available for comparison.
- The spiked and unspiked portions are otherwise treated comparably.
- The method explains what recovery criterion is suitable for this sample type and purpose.
- Other QC evidence, such as blanks and duplicates, is considered alongside the spike result.
- The public communication labels spike recovery as a method-quality check rather than a treatment outcome.
What Would Weaken It?
- The report shows “95% recovery” without saying what was added or recovered.
- The original sample result is missing.
- The spike is added at a level far outside the range relevant to the actual samples without explanation.
- The communication turns spike recovery into removal efficiency, product effectiveness or environmental cleanup.
- The result is used to claim the method is perfect across every kind of sample.
- A poor spike result is ignored while favourable sample results are still promoted without qualification.
Worked Case 1: The Filter Advertisement
A filter company’s fictional infographic says “95% recovery” beside laboratory data and then says “removes 95%”. If the 95% came from a spike-quality-control sample, the graphic has changed the scientific job. To support 95% removal, the experiment would need a before-and-after treatment comparison of the relevant quantity under defined conditions.
Worked Case 2: The Method With 70% Recovery
A method recovers only 70% of a known addition in a particular sample type. Does that automatically mean every reported result is 30% too low? Not necessarily. The relationship may be more complicated, and the method may include corrections, uncertainty or acceptance rules. The correct response is to investigate the method’s validation and QC interpretation rather than apply a made-up universal correction.
Worked Case 3: The 105% Recovery
A spiked sample gives 105% recovery. A pupil says, “That proves the test is better than perfect.” No. Measurement variation or interference can make the calculated recovery slightly above the known addition. The question is whether the value is consistent with the method’s expected performance and other QC evidence.
Worked Case 4: Recovery in Two Different Sample Types
The same method shows close-to-expected recovery in clean water but poorer recovery in a muddy sample. That difference may reveal that the surrounding sample material affects the measurement. It does not by itself prove why. Further method investigation is needed.
Tempting Reasoning That Fails
- “Recovery means removal.” In this QC context, recovery refers to a known added amount recovered by the measurement process.
- “95% recovery means the original sample contains 95% of something.” The percentage is relative to the spike addition, not automatically the original sample composition.
- “100% recovery proves the method has no error.” One QC result cannot rule out every error source.
- “More than 100% is physically impossible.” The calculated measurement recovery can exceed 100% because of measurement variation or interference.
- “A bad spike means the original sample result must simply be deleted.” It is evidence that the method’s performance for that sample needs evaluation, not permission to hide inconvenient data.
Model and Measurement Limits
Spike recovery is only one part of quality control. It does not replace blanks, duplicates, reference checks, calibration, sample handling or sampling design. It also depends on when and where the known amount is added. A spike introduced late in the laboratory process may test fewer steps than one carried through more of the preparation chain.
That timing matters because a QC check can only reveal failures that occur inside the part of the process it actually travels through.
How Far Can the Conclusion Travel?
A satisfactory spike recovery can strengthen confidence that the analytical method can account for a known added amount in that sample and procedure. It does not automatically establish the true concentration of every untested sample, the effectiveness of a treatment, the safety of a product, or the environmental importance of the measured substance.
One evidence object, one job. Broader conclusions require broader evidence.
PSLE-Style Transfer Case
A fictional test measures 40 units of Substance T in a sample. A second equal portion receives 10 known extra units. The spiked portion is measured as 49 units.
Question: What does the result suggest about recovery of the known addition?
Reasoned answer: The measurement increased by 9 units after 10 units were deliberately added. In this simplified case, the method recovered about 90% of the added amount. That is evidence about the method’s response to the spike. It does not mean 90% of the original Substance T was removed.
Explained Practice
Practice A: A report says “spike recovery 98%”. What is the first question? “98% of what known added amount?”
Practice B: An advertisement says a purifier removes 98% because a laboratory spike recovered 98%. What is missing? A separate treatment-removal experiment comparing the relevant amount before and after purification under controlled conditions.
Practice C: A spike recovery is poor only in one complex sample type. What alternative should stay alive? The surrounding sample material may be interfering with preparation or measurement.
Delayed Independent Return: The S-P-I-K-E Check
- S — Spike: What known amount was deliberately added?
- P — Pair: What did the original and spiked portions measure?
- I — Increase: How much extra did the method actually measure?
- K — Known baseline: Is the percentage relative to the known addition rather than the original sample?
- E — Evidence boundary: What method-quality claim does this support, and what treatment or product claims remain untested?
Parent and Tutor Teaching Guide
Use counters rather than laboratory jargon. Put 30 counters in one box to represent the original sample. Put 30 counters in a second box, then deliberately add 20 red counters. Imagine the measuring method notices 19 of the 20 red counters. Ask: “Did we remove 95% of the original 30?” The physical model makes the denominator mistake obvious.
Then change the headline. Write “95% spike recovery”, “95% removed” and “95% of samples positive” on three cards. Ask the learner to explain what denominator each phrase would require. The teaching goal is not technical chemistry; it is disciplined reading of percentages.
Authoritative Sources
- Singapore Examinations and Assessment Board — 2026 PSLE Science Syllabus
- Ministry of Education, Singapore — 2023 Primary Science Teaching and Learning Syllabus
- U.S. EPA — Quality Assurance Handbook and Toolkit for Participatory Science Projects
- U.S. Geological Survey — 2026 field methods, quality-assurance and data-management plan
EPA guidance describes spiked samples as samples containing a known added amount and uses percent recovery as a way to evaluate measurement performance. USGS field-quality programmes also use spikes alongside blanks and replicates. The official Singapore Primary Science and PSLE framework gives learners the reasoning tools needed to interpret such evidence: analyse information, evaluate methods, consider uncertainty and communicate a conclusion that does not outrun the data.
The Quiet Return
A percentage is never self-explanatory.
Its meaning lives in the experiment that created its numerator and denominator.
When science says “95% recovery”, do not ask only whether 95% is high. Ask what was known, what was added, what was measured, and what the percentage is actually about.
