PSLE Science Reality Lab Vol No.268 | “The Headline Says ‘Up 200%’” — How Big Was the Starting Amount?

PSLE-SCI-REALITY-0268

Wait, What? A 200% Increase Can Be Only Two More

A science headline says, “Detected particles increased by 200% after the change.” That sounds enormous. A learner pictures hundreds of extra particles appearing.

Then the original data are revealed. The first observation was 1 particle. The later observation was 3 particles.

The increase really is 200% relative to the starting value: the amount rose by 2, and 2 is twice the starting amount of 1. But the absolute increase was still only 2 particles.

Now compare a second experiment that rises from 100 particles to 120 particles. The percentage increase is only 20%, yet the absolute increase is 20 particles — ten times the absolute change in the first case.

Both percentage statements can be mathematically correct. They answer a different question from “How much did the amount actually change?”

The Reality Lab habit is: when a headline gives a large percentage change, recover the starting value before deciding how large the scientific change really was.

Quick Answer

Find the earlier or starting value. It is the baseline used in the percentage calculation.
Find the later value and calculate the absolute difference.
Keep two statements separate: absolute change tells how many units changed; percent change tells how large that change is relative to the starting value.
Check that the two measurements use the same variable, units, method, place, population and comparable time window.
Do not treat a large percentage as proof of a large real-world effect until you know the baseline and scale.
Do not use the ordinary percent-change formula when the starting value is zero; the usual calculation has no valid non-zero baseline to divide by.

The Exact Learner Job This Volume Owns

This volume owns one evidence-transfer problem: how to evaluate a real-world scientific headline, infographic, product comparison or chart that reports a dramatic percent increase while hiding or de-emphasising the starting amount.

It does not become the general owner of percentages, ratios, graph reading or mathematics. Existing PSLE Science and Mathematics pages retain those jobs. Reality Lab applies one mathematical relationship to the scientific communication problem that appears when “up 200%”, “down 80%” or “three times higher” is allowed to stand without the baseline that gives the percentage meaning.

Rebuild the Evidence Object: The Two Headlines

These are original constructed examples.

Case	Start	End	Absolute increase	Percent increase
A	1 unit	3 units	+2 units	+200%
B	100 units	120 units	+20 units	+20%
C	50 units	75 units	+25 units	+50%
D	500 units	550 units	+50 units	+10%

If a headline shows only the rightmost column, Case A looks like the biggest change. If the scientific job is to know how many units were added, Case D changed the most. Neither reading makes the other false. The mistake comes from answering one question with a number designed for another.

The Calculation Is Simple; the Evidence Job Is Not

The U.S. Bureau of Labor Statistics explains percent change by subtracting the earlier value from the later value, dividing the difference by the earlier value and multiplying by 100. The earlier value therefore sits inside the meaning of the result.

For an increase from 40 to 60:

Absolute increase = 60 − 40 = 20
Percent increase = 20 ÷ 40 × 100 = 50%

The denominator is not background decoration. It is the starting point that makes “50%” mean what it means.

Observed, Calculated, Claimed and Inferred

Layer	Example
Observed	The measured value was 1 unit before and 3 units after.
Calculated	The absolute increase is 2 units; the relative increase is 200%.
Claim	The measured quantity rose by 200% over this comparison.
Unsupported inference	The change must be scientifically huge, important, harmful, beneficial or larger than a different change whose baseline is much bigger.

The calculation can be perfectly correct while the interpretation still outruns the evidence.

Why Small Baselines Produce Big Percentages

A percentage change asks how large the difference is compared with the starting value. When the starting value is small, even a small absolute difference can be large relative to that start.

Suppose a detector records 2 events in Week 1 and 6 in Week 2. The absolute increase is 4 events. Relative to the starting value of 2, that increase is 200%.

Now suppose a second detector records 1,000 events and later 1,100. Its absolute increase is 100 events, but its percent increase is only 10%.

The first case has the larger percentage. The second has the larger absolute increase. A responsible science headline should make clear which kind of magnitude matters to the claim.

The Baseline Is Part of the Scientific Claim

A baseline can be a measurement taken before an intervention, an earlier year, a control condition, a reference sample or another defined starting point. Choosing a different baseline can change the reported percentage even when the later value stays the same.

Imagine a river measurement that ends at 60 units:

Chosen start	End	Percent change
30	60	+100%
40	60	+50%
50	60	+20%

The endpoint did not change. The percentage did because the baseline changed. This is why phrases such as “compared with last week”, “compared with the 2010–2020 average” or “compared with the untreated sample” are scientifically important.

The Representation Check: What Did the Headline Leave Out?

Percentages are attractive headline material because they compress two numbers into one dramatic-looking number. But compression removes information. A headline saying “UP 300%” may leave out:

the starting amount;
the ending amount;
the unit;
the number of observations behind each value;
the time interval;
measurement uncertainty;
whether the measurement method changed;
whether the compared groups are actually comparable.

A useful percentage is not suspicious merely because it is compressed. It becomes scientifically weak when the missing context prevents the reader from reconstructing the evidence.

The Comparison Check: Are the Start and End Measuring the Same Thing?

Before calculating or trusting a percent change, verify the comparison. A perfectly executed percentage calculation can still compare unlike quantities.

Are the units the same?
Was the same instrument or method used?
Was the sampling area the same?
Were the observation durations equal?
Was the same age group, species, material or product category measured?
Did a reporting rule or detection threshold change?
Was one value a count while the other was a rate?

If the measurement definition changes, “percent change” can partly measure the changed method rather than the changed world.

The Denominator Check: What Exactly Was the Percentage “Of”?

Students often treat a percentage as if it floats freely. It does not. A percent always depends on a denominator or reference whole.

“200% increase in detections” may mean the difference divided by the earlier number of detections. “30% of samples positive” uses number of samples as the denominator. “20% concentration reduction” uses the initial concentration. These percentages cannot be swapped merely because they share the percent symbol.

The Reality Lab question is: 200% of what starting amount, and measured how?

The Zero-Baseline Problem

What if the starting value is zero and the later value is 5? The ordinary percentage-change calculation would require division by zero. That is not a valid ordinary percent increase.

A careful communicator should report the absolute change — for example, “from 0 detections to 5 detections” — and explain the measurement context rather than inventing an enormous percentage.

This becomes especially important when a detection threshold is involved. “Zero” can mean true zero, below detection, none observed, none recorded or a rounded value. Reality Lab already has separate owners for zero, blank, non-detect and reporting limits. This volume does not re-own those jobs.

Percentage Points Are Not Percent Change

Another communication trap appears when the measured quantity is itself a percentage.

If germination rises from 20% to 30%, the direct difference is 10 percentage points. Relative to the starting 20%, the increase is 50%. Both statements can be mathematically correct, but they describe the change differently.

Writing “up 10%” when the writer means “up 10 percentage points” can mislead readers. A careful student checks whether the original measurement is already expressed as a percentage before interpreting the change.

A Large Percentage Is Not Automatically a Large Scientific Effect

Scientific importance depends on more than relative change. A 300% increase in an extremely rare event may still leave the event rare. A 5% change in a very large flow, population or energy use may correspond to a huge absolute amount.

Importance can also depend on thresholds, biological mechanisms, engineering tolerances, uncertainty, costs or consequences. Reality Lab does not decide those specialist questions from the percentage alone. It teaches the learner to preserve the measurement scale so that the appropriate specialist question can be asked.

Worked Case 1: The Three Colonies

A fictional dish shows 1 visible colony under Condition A and 3 under Condition B. A caption says, “Colony count increased by 200%.”

The percent calculation is correct if those are comparable counts. But the learner should also report that the absolute difference is 2 colonies and ask about replication, countability, method and whether one dish represents a wider process. A dramatic percentage does not repair weak sampling.

Worked Case 2: The Two Filters

Filter X reduces a constructed particle reading from 100 to 50 units. Filter Y reduces another test from 4 to 1 unit.

Filter X has an absolute reduction of 50 and a relative reduction of 50%. Filter Y has an absolute reduction of 3 and a relative reduction of 75%. Which is “better”?

The percentage alone cannot answer. Were starting conditions comparable? Are the units, particle types, flow rates and test durations the same? Which outcome matters? A relative reduction is one piece of the evidence, not a universal ranking machine.

Worked Case 3: The Tiny Baseline Headline

A news-style graphic says, “Sightings rose 400%!” The small print shows that monthly sightings rose from 1 to 5 after survey effort expanded from 10 hours to 80 hours.

Two issues now matter. First, the 400% increase starts from only 1 sighting. Second, observation effort changed eightfold. The percentage of raw sightings therefore cannot be read as a clean change in animal abundance. Reality Lab Vol No.060 owns the search-effort problem; this volume uses it as a boundary reminder.

Worked Case 4: Same Absolute Increase, Different Percentages

Sensor A rises from 10 to 20 units. Sensor B rises from 100 to 110 units. Both increase by 10 units.

Sensor A increases 100%. Sensor B increases 10%. The absolute changes match, but the relative changes do not. If the scientific question asks “Which changed by more units?”, they tie. If it asks “Which changed by a larger fraction of its starting value?”, A is larger.

Worked Case 5: Different Absolute Changes, Same Percentage

Plant A grows from 10 cm to 15 cm. Plant B grows from 100 cm to 150 cm. Both increase 50% relative to their own starting heights.

Yet Plant A gains 5 cm and Plant B gains 50 cm. Same relative change; different absolute change. The comparison question determines which representation is informative.

Worked Case 6: Method Change Creates a False Trend

A detector recorded 40 events last year and 80 this year. The headline says “events doubled”. Later, the methods reveal that the old detector could register events only above 10 units while the new detector detects events above 5 units.

The arithmetic percentage is still 100%, but the scientific comparison is confounded by a changed detection rule. Some of the apparent increase may come from the measurement system. The claim needs method continuity or a correction that makes the two periods comparable.

What Evidence Strengthens a Percent-Change Claim?

the starting and ending values are shown, not only the percentage;
units and measurement definitions match;
the time windows are comparable;
sampling effort is similar or appropriately adjusted;
instrument and detection methods are consistent;
replicate measurements show that the pattern is not one unstable result;
uncertainty is small enough for the claimed change to be distinguishable;
absolute as well as relative magnitude is reported when it matters;
the chosen baseline is scientifically justified rather than selected only because it creates a dramatic percentage.

What Weakens It?

the headline gives “up 300%” without a starting value;
the comparison begins from an unusually tiny or cherry-picked baseline;
the measurement method changes between the two values;
the first value is zero but a normal percentage increase is still claimed;
percentage points and percent change are mixed;
different units or populations are compared;
a large percentage is presented as proof of a large practical or biological effect without showing scale;
the result depends on one observation with no replication.

Tempting Reasoning That Fails

“200% means 200 extra units.” No. It is relative to the baseline.
“The largest percentage must be the largest absolute change.” Small baselines can create large percentages.
“A 10% change is always small.” Ten percent of a very large quantity can be enormous.
“A 300% increase proves the intervention had a huge effect.” First establish comparability, baseline, uncertainty and causation.
“From 0 to 5 is an infinite percent increase.” The ordinary percent-change calculation divides by the starting value and is not defined when that value is zero.
“20% to 30% is a 10% increase.” It is 10 percentage points, or a 50% relative increase from the original 20%.
“If the maths is correct, the scientific claim is correct.” Correct arithmetic cannot fix an unfair or mismatched comparison.

Alternative Explanations Before You Credit the Change

A measured increase can have several explanations besides the one promoted in a headline. The environment may have changed. Sampling effort may have increased. The instrument may be more sensitive. The starting value may have been unusually low. The population sampled may differ. Random variation may matter. An intervention may truly have caused the increase — but causal credit needs its own evidence.

Reality Lab therefore keeps two questions apart:

How large is the numerical change?
Why did the change occur?

A percent-change calculation addresses the first question only partly. It does not answer the second.

How Far Can the Conclusion Travel?

If two comparable measurements rise from 10 to 30 units, the learner can responsibly report an absolute increase of 20 units and a relative increase of 200% from the starting value of 10.

The learner cannot automatically conclude that the change is important, beneficial, harmful, caused by a particular factor or larger than every lower-percentage change. Those conclusions require the scale, uncertainty, mechanism and decision context.

PSLE-Style Transfer Case

Two trays are observed in an original constructed investigation.

	Before	After
Tray P	2 seedlings	6 seedlings
Tray Q	40 seedlings	60 seedlings

A pupil says, “Tray P had the bigger scientific increase because 200% is bigger than 50%.”

Reasoned response: Tray P has the larger relative increase: it rose by 4 seedlings, which is 200% of its starting value of 2. Tray Q has the larger absolute increase: it rose by 20 seedlings, or 50% of its starting value of 40. Which change is more important depends on the investigation question and whether the trays were fairly comparable. The percentage alone does not settle that.

Delayed Independent Return: The B-A-S-E Check

B — Baseline: What was the starting value?
A — Absolute change: How many actual units changed?
S — Same measurement? Are method, units, population and time window comparable?
E — Effect claim: Does the percentage really support the importance or cause being claimed?

Explained Practice

Practice A: A count rises from 5 to 15. Absolute increase? 10. Percent increase? 200% relative to the starting 5.

Practice B: A count rises from 500 to 550. Absolute increase? 50. Percent increase? 10%. This has a larger absolute increase than Practice A despite the smaller percentage.

Practice C: A rate rises from 0% to 5%. Can ordinary percent change be calculated from zero? No. Report the change directly and preserve the denominator and context.

Practice D: Efficiency rises from 40% to 50%. Difference? 10 percentage points. Relative increase from the original efficiency? 25%.

Practice E: Detections rise 100%, but survey hours also double. What must be checked? Detection rate or another effort-adjusted measure before claiming the underlying phenomenon doubled.

Routes to Existing Canonical Owners

Parent and Tutor Teaching Guide

Start with physical counters. Put 1 counter in one group and 100 in another. Add 2 counters to the first and 20 to the second. Ask two different questions: “Which group had the larger percent increase?” and “Which group gained more counters?” The learner should discover that the answers differ.

Next, hide the starting numbers and show only “+200%” and “+20%”. Ask which change is larger. Most learners will choose 200%. Reveal the baselines and let them revise the answer. This makes the missing-baseline problem memorable without teaching distrust of percentages.

Then change only the baseline while keeping the endpoint fixed. For example, end at 60 after starting at 30, 40 and 50. The learner sees that percentage change belongs to a relationship between two values, not to the endpoint alone.

Finally, introduce one unfair comparison: double the observation time in the second condition. Ask whether the percent calculation can be arithmetically correct but scientifically misleading. This reconnects Mathematics to scientific inquiry.

Authoritative Sources

The official Singapore Science frame asks learners to interpret and analyse information, evaluate observations and information, identify assumptions and uncertainty, and communicate explanations based on evidence. A percent-change headline is an ideal transfer object because the arithmetic may be easy while the evidence boundary is not.

The Quiet Return

A percentage is not a magnifying glass that tells you how important a scientific result is.

It is a relationship to a starting value.

When the headline shouts “UP 200%”, look for the quiet number it started from.

Small Group Tutorials