Did you know that a graph can look perfectly smooth even when a sensor missed part of an experiment? A student takes a temperature reading, saves a chart and confidently explains a sudden change. But a blank patch in the data may have been hidden when the software joined two points with a line. Three small words can prevent a big misunderstanding: timestamp, sampling interval and missing.
The useful answer is to check when readings were recorded, how often the sensor was supposed to record them and whether any readings are absent before interpreting a graph. A timestamp locates an observation in time, a sampling interval is the time between observations, and a missing value means the dataset does not contain a usable observation at that point. These facts matter as much as the measured numbers.
A fictional cup-of-water investigation
Picture a secondary Science class testing how the temperature of water changes after a warm cup is placed on a desk. A digital sensor is set to record once every five seconds. The first readings are taken at 0, 5, 10, 15 and 20 seconds. At 25 seconds the connection drops. The next successful reading appears at 35 seconds. The software draws a line between the 20-second and 35-second points.
If a learner says, ‘The temperature changed steadily at 25 and 30 seconds,’ the learner is claiming more than the data show. The line may be a visual connection or an interpolation, not a measurement at those missing times. This is a fictional classroom example, not a report of real experimental results.
Name the time column correctly
Timestamp: when that row belongs
A timestamp is a recorded point in time. It might look like ’10:32:15′ for a clock time or ’15 s’ as elapsed time since an experiment started. These formats are not the same. A wall-clock timestamp can describe an actual moment on a clock; elapsed time states the duration since a reference event. Always check the header and units before subtracting values.
The word family includes time, timestamp and time-stamped. In laboratory writing, ‘The reading was time-stamped at 10:32:15’ is different from ‘The reading was recorded 15 seconds after the timer started’. Precision matters when comparing two sensors, since their clocks may not have been synchronised.
Sampling interval: the gap between successive observations
If readings are collected at 0, 5, 10 and 15 seconds, the sampling interval is five seconds. The sampling rate is the reciprocal: one reading per five seconds, or 0.2 samples per second (0.2 Hz). A learner should never confuse ‘five seconds per sample’ with ‘five samples per second’; those would describe very different data density.
In its sensor hardware information, Vernier describes capabilities in samples per second. The maker’s Graphical Analysis user manual separately defines rate and interval as reciprocal quantities and describes time-based collection at regular intervals. Devices have different limits and settings; the purpose is to read the labels in the particular software, not to memorise one default sampling rate.
Missing: not recorded, not zero
A blank, a dash, NA or an absent row can mean an observation is unavailable. It does not automatically mean the sensor measured a value of zero. A thermometer recording no data at 25 seconds does not establish a water temperature of 0°C. Equally, an instrument displaying zero does not always mean the physical quantity vanished; it may reflect calibration, rounding or a sensor reading. Context decides.
Useful related words are record, recorded, unrecorded, dropout, gap, interpolate and estimate. Interpolate means estimate an unknown value between known values using a method. An interpolated point must not be labelled as if a sensor actually measured it. Extrapolation goes beyond the observed range and introduces different uncertainty.
Read the five-second log together
- 0 seconds: 48.0°C, recorded.
- 5 seconds: 46.7°C, recorded.
- 10 seconds: 45.8°C, recorded.
- 15 seconds: 44.9°C, recorded.
- 20 seconds: 44.2°C, recorded.
- 25 seconds: no reading available.
- 30 seconds: no reading available.
- 35 seconds: 42.1°C, recorded.
Every value in this illustrative log is invented solely to teach interpretation. There are eight scheduled timestamps from 0 through 35 seconds inclusive, but only six recorded temperature values. The programme may show a line between 44.2°C and 42.1°C; it must not quietly turn that drawn line into two extra verified readings. A useful description is: ‘Readings are missing at 25 and 30 seconds because of a reported connection interruption; the observed temperature fell from 44.2°C at 20 seconds to 42.1°C at 35 seconds.’
Mathematically, the measured difference across those two recorded endpoints is 2.1°C over 15 seconds. If a student calculates an average change over that gap, they must label it as an average between endpoints, not proof that the temperature fell at a constant rate at every instant. Even with a dense dataset, a measured average and an instantaneous change are different concepts.
Why the graph alone can mislead
Line charts often connect points to help the eye follow a series. Connecting two valid endpoints does not tell us what happened between them. If the sampling interval changes or the sensor disconnects, equal spacing on a screen can hide unequal time gaps unless the horizontal axis uses actual timestamps. Check whether the chart is a time-series plot with a numerical time axis, not a category plot treating rows as evenly spaced labels.
Another trap is the phrase ‘a reading every five seconds for 30 seconds equals six readings’. Depending on whether the device counts an initial reading at zero and a final reading at 30 seconds, the count may be seven, six or something else. The actual software setup and recorded timestamps settle the count. Don’t invent a fixed endpoint convention.
A classroom method that teaches the idea, not the vocabulary list
Ask learners to hand-record five observations at known intervals, then intentionally leave one row blank on a copy of the table. Do not damage the original data. Have one student describe the gap using everyday words and another use the terms ‘sampling interval’, ‘missing observation’ and ‘estimated value’. Compare their claims. The better explanation distinguishes recorded, absent and inferred data.
For a Primary child, begin with a paper weather chart: ‘Monday observed, Tuesday blank, Wednesday observed.’ A blank Tuesday is not zero rainfall unless a reliable record actually says 0 mm. Lower-secondary learners can calculate intervals. G1, G2 and G3 learners can work with age-appropriate datasets without assuming their syllabuses ask the same tasks. Older students can discuss sensor calibration and uncertainty.
For active recall, cover the timestamp column and ask the learner to predict it for regularly sampled data; then reveal a missing interval and ask what changed. Tomorrow, present readings at 0, 2, 4, 10 and 12 seconds and see whether the learner identifies the six-second gap. Transfer the same thinking to exercise-tracker gaps, traffic sensor logs or study-app activity reports.
Misconceptions to correct immediately
- Wrong: ‘Missing means zero.’ Correction: Missing indicates no usable recorded observation; zero is a numerical value.
- Wrong: ‘2 Hz means one reading every two seconds.’ Correction: 2 Hz means two readings per second, an interval of 0.5 seconds.
- Wrong: ‘A straight line proves the process was constant.’ Correction: A plotted line may simply connect measurements.
- Wrong: ‘The tenth row is at ten seconds.’ Correction: That only follows if the timestamp rules and row numbering support it.
Use the words to improve English explanations and research habits
Strong English sentences often distinguish an observation from a conclusion. Compare ‘The water cooled smoothly’ with ‘The observed readings decreased, but two intervals were unrecorded’. The second sentence is more careful about evidence. In Science, this is intellectual honesty. In Mathematics, it is respect for units and axes. In daily life, the same distinction helps families read sleep trackers or download progress graphs without treating every displayed point as a direct observation.
The eduKateSG guide to sensors, feedback and robotics provides the broader mechanism. This article solves the narrower reader problem of interpreting time columns and gaps. The vocabulary hub offers additional ways to turn words into reasoning.
Check your understanding
- A sensor takes readings at t = 0, 4, 8 and 12 seconds. What is its sampling interval and rate?
- A temperature series shows 19°C, blank, 20°C. Can you claim the middle reading was 0°C?
- If a chart connects a reading at 10 seconds to one at 40 seconds, what should you check before describing the intervening trend?
- Which statement is better: ‘The sensor proved the temperature was stable at 30 seconds’ or ‘No observation was recorded at 30 seconds’?
Answers: four seconds, or 0.25 samples per second; no; check recorded timestamps, missing observations and interpolation; the second is supported if the record really contains no measurement. For deeper practice, write a one-sentence conclusion that clearly separates measured facts from an estimate.
Frequently asked questions
Are an interval and a duration the same?
No. The sampling interval is the gap between successive readings. Duration is the total time over which collection occurs. A 60-second recording can use 1-second, 5-second or other intervals, producing different numbers of expected readings.
Should students delete missing rows before plotting?
Not automatically. Deleting gaps can conceal important problems. Preserve the original dataset; decide whether the plotting tool should show gaps, explicit missing markers or method-labelled estimates. State any cleaning or interpolation in the report.
Does a high sampling rate always produce better evidence?
No. More samples can help capture fast changes, but sensor accuracy, noise, the phenomenon, memory limits and good experimental design matter too. Choose settings appropriate to the question and instrument, not simply the largest number on the screen.
A habit for the next experiment
Before writing what happened, ask: ‘When was this measured? How far apart were the readings? Which points are missing or estimated?’ Those three questions turn technical vocabulary into stronger evidence, clearer graphs and more trustworthy school explanations.