Q: What does Measurement and Uncertainty for IP Physics: From Sig-Figs to Spreadsheet Graphs cover? A: How to align IP Physics measurement and uncertainty skills to SEAB's 2026 syllabuses, including sig-fig rules,
uncertainty propagation, and spreadsheet-based analysis.
TL;DR - New syllabus = new marks. Quote instruments to the correct significant figures, report uncertainties as ±21 least-division, and round calculated quantities to the smallest-sf raw datum. Master absolute → fractional → percentage error conversions; propagate by addition for x / ÷, by sum of percentages for powers. In H2 Physics (9478), Paper 4 (Practical) can include spreadsheet-based data processing (graphs, trendlines, gradients and areas). Try the 20-s pendulum timing lab below; analyse your data with Sheets/Excel and you will have practised every assessed skill in one evening.
Need more structured drills? Follow the weekly diagnostics inside our IP Physics hub
so measurement practice lines up with the rest of your upper-sec revision.
1 Why measurement & uncertainty just became high-stakes
In the 2026 H2 Physics (9478) syllabus, Paper 4 (Practical) can include data analysis where candidates are required to process and analyse data using spreadsheet software.
At O-Level Physics (6091), practical work still rewards accurate measurement, correct units and appropriate significant figures - the syllabus explicitly warns that misuse of units/significant figures is liable to be penalised.
2 Reading instruments & recording to the right sig-figs
2.1 Analogue versus digital
Instrument
Least division / resolution
Raw reading rule
Example
Metre rule
1mm
±0.5mm
42.0cm±0.05cm
Thermometer
1∘C
±0.5∘C
23.0∘C±0.5∘C
Digital stopwatch
0.01s
quote all decimals
12.37s
SEAB’s syllabuses emphasise correct units and appropriate significant figures. In practice, match your final rounding to the precision of your raw data and keep intermediate rounding conservative.
For example, calculating R=V/I from 3 s.f. voltage and 2 s.f. current means your final R must be 2 s.f.
2.2 Sig-fig decision tree
Quote what the instrument shows - no truncation.
Intermediate steps: keep 1 extra sig-fig to suppress rounding drift.
Final line: match the least sig-fig among the raw inputs.
Worked example I=0.43A (2 s.f.), V=3.05V (3 s.f.) R=7.09Ω (calculator) → round to 2 s.f. →7.1Ω.
3 Absolute, fractional & percentage uncertainty
Symbol
Definition
Δx
Absolute uncertainty (same unit as x)
xΔx
Fractional uncertainty (no units)
xΔx×100%
Percentage uncertainty %
Rule-of-thumb: when you add/subtract, add absolute uncertainties; when you multiply/divide or raise to a power n, add percentage uncertainties and multiply by n.
Use this route before substituting numbers:
Calculation type
Uncertainty form to use first
Why this form fits
Final check
Direct reading such as length or time
Absolute uncertainty
It stays in the same unit as the reading.
Does the uncertainty match the instrument resolution?
Addition or subtraction
Absolute uncertainties
The largest absolute spread controls the final sum or difference.
Are all terms in the same unit before adding?
Multiplication or division
Percentage uncertainties
Relative spreads combine when quantities scale each other.
Convert back to absolute uncertainty only at the end.
Powers such as squared period
Percentage uncertainty multiplied by the power
A power amplifies relative uncertainty.
Did you multiply the percentage uncertainty before rounding?
Before choosing a propagation rule, identify where the uncertainty came from:
Uncertainty source checkpoint
First move
What to record
Common trap
Single analogue reading
Use half the smallest division.
Reading plus absolute uncertainty, in the same unit.
Quoting too many decimal places because the calculator can.
Single digital reading
Use the display resolution unless the task states otherwise.
Reading with all displayed decimals.
Dropping a displayed zero because it "looks unnecessary".
Repeated readings
Compare the spread before using the mean.
Mean value plus a scatter estimate such as half range.
Averaging first, then pretending the spread disappeared.
Zero error or offset
Correct every raw reading before later calculations.
Corrected value plus the remaining uncertainty.
Treating a shifted scale as random scatter.
Worked check: five timings for 20 oscillations are 31.18,31.24,31.20,31.26,31.22s. The mean is 31.22s, the range is 0.08s, so a simple scatter estimate is half range: ±0.04s. Only after that should you divide by 20 to get the period and its timing uncertainty.
Misconception to avoid: repeated readings reduce random scatter, but they do not erase instrument resolution or a zero error. Name the source before you propagate the number.
Quick drill
A meter measures L=0.381±0.002m. Percentage uncertainty =0.3810.002×100=0.52%.
Community examples show the same propagation applied to pendulum g calculations.
4 Spreadsheet skills now examinable
SEAB’s H2 Physics (9478) syllabus includes spreadsheet-based data processing as part of Paper 4 (Practical), and lists core tasks like plotting graphs, adding a trendline and displaying its equation, and determining gradients/areas from data.
Practically, that means you should know how to:
Import a .csv file or type raw readings.
Use formulae and copy them correctly across cells (e.g. compute T=t/20, T2, lnT).
Plot a labelled line graph, adjust axis scales, and add a line of best fit using built-in tools.
Display the trendline equation and extract the gradient/intercept where appropriate.
Use numerical methods (e.g. trapezium/summation ideas) to estimate areas and gradients from data when required.
5 Mini-Lab: Timing a pendulum with your phone & Sheets
This 30-minute activity rehearses every skill the syllabus demands.
5.1 Setup
String, small metal nut, metre rule, phone timer/camera.
Measure L three times; take the mean length. Record to the nearest millimetre.
5.2 Data collection
Displace under 15° (small-angle).
Start timing on the first crossing; record 20 oscillations to reduce percentage timing error.
Repeat x 5 trials.
Trial
t20 / s
T=t20/20 / s
1
31.20
1.560
…
…
…
Reaction time of 0.2 s roughly halves when timing 20 periods versus one, cutting % uncertainty dramatically.
5.3 Spreadsheet crunch
In Sheets, enter L and corresponding T.
Create columns for T2 and lnT if you wish to test alternative models.
Plot T2 vs L; gradient m=g4π2.
Use =LINEST(T2_range, L_range, TRUE, TRUE) to obtain m and its standard error; convert to % uncertainty.
