PSLE-SCI-REALITY-0139
Wait, What? An Empty Box Can Mean “We Do Not Know,” Not “There Is Nothing There”
A scientific map shows measurements across a region. Most grid cells contain numbers. One cell is blank.
A reader says, “That area must have zero.”
That conclusion can be completely wrong. A blank scientific value may mean the sensor did not report, the observation failed a quality check, the satellite was blocked by cloud, the sample was not collected, the result was withheld because it was invalid, or the data have not yet been processed. A measured zero is different: it is an observation whose reported value is zero under a defined method.
The Reality Lab habit is: never let empty space silently become a scientific measurement.
Quick Answer
- Check the legend, footnote or data dictionary for the meaning of blank, NA, null, missing or flagged values.
- Separate measured zero from no valid measurement.
- Ask why the value is missing: no observation, sensor failure, quality rejection, cloud cover, maintenance, processing delay or another cause.
- Do not include missing values as zero in totals or averages unless the scientific method explicitly justifies that treatment.
- Keep the conclusion narrow: a blank cell tells you about the evidence record, not automatically about the physical system.
The Exact Learner Job This Page Owns
This page owns one real-world evidence-transfer job: evaluating a scientific map, dashboard or table in which a blank or missing value is visually mistaken for a physical zero.
It does not replace the canonical owners for graphs, tables, missing measurements, detection limits or data quality. It applies those skills to a communication problem created by the representation itself: the display contains nothing, so the reader imagines that the measured world contained nothing.
- Reality Lab Vol No.039: “There Are No Gaps on the Map” — Were Some Values Filled In?
- Reality Lab Vol No.040: “The Pixel Has a Number” — What Does Its Quality Flag Say?
- Reality Lab Vol No.130: “Below the Detection Limit” — Is It Scientific to Enter Zero?
- How to Turn Raw PSLE Science Observations Into a Results Table Without Mixing the Variables
Original Reality Lab Case: The Blank Lake Cell
This is an original composite case with constructed data.
A monitoring map shows fictional Indicator M at nine locations around a lake.
| Location | Reported Indicator M | Data note |
|---|---|---|
| A | 4 | Valid |
| B | 5 | Valid |
| C | 0 | Valid measured zero |
| D | — | Sensor offline |
| E | 7 | Valid |
Locations C and D may look similar on a hurried chart if both are shown without colour. Scientifically they are different. At C, the instrument produced a valid value of zero under the method. At D, there is no valid observation for that time.
If a pupil replaces D with zero, the data record has been changed. The pupil has created an observation that was never made.
Observed, Missing and Inferred
| Display state | What we can say | What we cannot assume |
|---|---|---|
| 0 | A valid result was reported as zero on the stated basis | That the physical quantity is absolutely absent under every possible method |
| Blank | No displayed valid value is available | That the true value is zero |
| Below detection | The method did not distinguish a signal above its stated capability | That the amount is exactly zero |
| Invalid/flagged | A value may exist but failed a stated quality rule | That the underlying physical phenomenon did not occur |
Why Missing Data Happen
Missingness has causes. Those causes can themselves matter scientifically.
- No observation: nobody sampled that place or time.
- Instrument failure: power, communication or sensor problems prevented a usable result.
- Quality rejection: a measurement was collected but failed a quality criterion.
- Obstruction: cloud, smoke, ice, shadow or another physical condition prevented remote sensing.
- Maintenance: equipment was removed or being calibrated.
- Processing delay: the observation exists but has not yet been released.
- Outside scope: the dataset was never intended to cover that location, time or condition.
Notice how different these explanations are. Missing data are not one physical state.
The Representation Check: What Symbol Was Chosen for Missingness?
Scientific displays use many conventions: blank cells, dashes, grey pixels, “NA”, “N/A”, “null”, special codes or cross-hatching. The same symbol can mean different things in different datasets.
A strong reader therefore does not guess from appearance. The reader checks the legend or metadata.
USGS datasets provide useful examples of quality flags and missing values. In some systems, a blank or null flag indicates that data quality information itself is missing; in others, measurements that fail serious checks may be transmitted as missing values rather than as ordinary numbers. The general lesson is durable: the data code must be read before the physical world is inferred.
The Arithmetic Trap: A Missing Value Can Change an Average Even Before You Notice It
Suppose four valid measurements are 4, 6, 8 and 10. A fifth observation is missing.
The average of the four observed values is 7. If someone enters zero for the missing fifth value, the new average becomes 5.6.
The arithmetic is correct for the invented list 4, 6, 8, 10, 0. The science is wrong because the zero was never observed.
Missing at Random? Sometimes the Gap Is Telling You Something
Imagine a weather sensor that fails only during the most intense storms. The missing periods are then not ordinary random gaps. If a report calculates an average using only the calm periods that survived, the result can become biased toward calmer conditions.
Primary learners do not need advanced missing-data statistics to understand the scientific habit: ask whether the reason for missingness is related to the thing being measured.
Blank on a Map Is Not Blank in Nature
A blank satellite pixel does not mean that nothing existed on the ground. It means the dataset did not provide a valid value there for the requested observation. A missing biodiversity record does not prove that a species was absent. A blank temperature reading does not mean absolute zero. A missing concentration result does not prove zero concentration.
The screen and the world are different objects.
What Evidence Would Strengthen the Interpretation?
- A legend that distinguishes zero, missing, invalid and below-detection values.
- Quality flags explaining why observations were withheld.
- Sensor or sampling logs showing when data collection failed.
- Neighbouring observations or independent methods used carefully to assess what may have happened during the gap.
- Transparent statements describing whether missing values were excluded, estimated or imputed.
- A sensitivity check showing whether conclusions change under reasonable treatments of missing data.
What Would Weaken a Claim Built on Blank Cells?
- The report silently treats blanks as zero.
- The legend does not define missing-value symbols.
- Missingness is concentrated exactly when conditions are most extreme.
- Invalid measurements are removed without reporting how many were removed.
- Map colours make “no data” visually identical to the lowest measured category.
- A claim of absence is made from a dataset with no valid observation at the relevant place or time.
Worked Case 1: The Bird Map
A citizen-science map has no bird record in one square. Can we say there were zero birds? Not unless the survey effort and detection process support that conclusion. The square may simply not have been surveyed.
Worked Case 2: The Cloudy Satellite Image
A vegetation map contains grey cells after heavy cloud. Grey can mean that no reliable surface observation was available. It does not mean vegetation index = 0.
Worked Case 3: The Broken Thermometer
A temperature logger records 28°C, 29°C, blank, blank, 30°C. If the sensor lost power during the two gaps, inserting 0°C would create a dramatic cooling event that never occurred in the evidence.
Worked Case 4: The “No Pollution Here” Headline
A map has a blank monitoring station and a headline says “zero pollution detected”. If no valid observation exists, the evidence supports “no valid result available”, not “zero pollution”.
Tempting Reasoning That Fails
- “Blank means zero.” A blank is a display state, not a measured quantity.
- “No record means absence.” Absence requires suitable observation effort.
- “If data were invalid, the phenomenon was not real.” Invalid measurement and physical absence are different claims.
- “Replacing missing values with zero is conservative.” It can bias results upward or downward depending on the question.
- “A smooth filled map is always better than one with gaps.” Filling gaps adds a model or interpolation layer that must be disclosed.
Model and Measurement Limits
Sometimes scientists estimate missing values from nearby observations, physical models or statistical methods. That can be useful, but an estimated value is not the same evidence object as a direct observation. Reality Lab Vol No.039 owns the specific problem of filled gaps; this article stops earlier at the first question: did a blank mean zero in the first place?
A measured zero also has limits. It is zero on the reporting basis of that method, not proof that absolutely no amount exists at any scale or under any more sensitive method.
How Far Can the Conclusion Travel?
From a blank cell we can often conclude only that the displayed dataset lacks a valid value there. To say what happened physically, we need additional evidence. The gap may later be resolved by a repaired sensor, another station, a second instrument, field observations or a model—but those are new evidence steps.
PSLE-Style Transfer Case
A data logger records dissolved oxygen every hour. At 2 p.m. and 3 p.m., the entries are blank because the probe was being cleaned. A pupil enters 0 mg/L for both hours before drawing the graph.
Question: Why is the graph scientifically misleading?
Reasoned answer: No valid oxygen measurements were made during those hours. Entering zero changes missing observations into measured zeros and creates an unsupported drop in the graph.
Explained Practice
Practice A: A table shows “—” and the legend says “not sampled”. What should you record? Missing/not sampled, not zero.
Practice B: A map uses white for zero and grey for no data. Why does the colour key matter? Because visually empty-looking regions represent different evidence states.
Practice C: A sensor fails only above a very high temperature. What danger appears if failed observations are ignored? The dataset may systematically miss the hottest conditions.
Delayed Independent Return: The B-L-A-N-K Check
- B — Basis: What symbol or code is being displayed?
- L — Legend: How does the dataset define it?
- A — Actual observation: Was a valid measurement made?
- N — Null reason: Why is the value missing or rejected?
- K — Keep separate: Do not turn missing, invalid, below-detection and zero into one category.
Parent and Tutor Teaching Guide
Draw five boxes and place 4, 5, 0, blank and “below detection” inside them. Ask the learner which boxes contain actual numerical observations and which contain information about the measurement process. Then ask what would go wrong if every non-number were replaced by zero.
For a second exercise, deliberately make “zero” and “no data” the same colour on a fictional map. Ask the learner to redesign the legend so the evidence states cannot be confused. This turns visual literacy into scientific reasoning rather than generic design criticism.
Authoritative Sources
- Singapore Examinations and Assessment Board — 2026 PSLE Science Syllabus
- Ministry of Education, Singapore — 2023 Primary Science Teaching and Learning Syllabus
- U.S. Geological Survey — Water Quality Chemical QA/QC Flags
- U.S. Geological Survey — Why Real-Time Streamflow Data May Be Revised
The official Singapore Science frame asks learners to interpret information, evaluate observations and methods, consider uncertainty and communicate reasoning. A blank data cell is a small but powerful test of those habits because the temptation is to infer physical reality from a display convention.
The Quiet Return
Zero is a result.
Blank is a question about the evidence record.
When the data display shows nothing, do not assume nature showed nothing too.
