⚗️ Experimental Physics & Lab Techniques
A measurement without an uncertainty is not a measurement. This is the craft of coaxing trustworthy numbers out of a noisy world — random vs systematic error, the √N law, error propagation, calibratio
What you’ll learn
- The Experimentalist’s CreedEstablish the course’s creed: every measurement is a value plus an uncertainty, and the uncertainty determines what you may conclude.A number without an uncertainty is meaningless; the ± is where the physics and the discovery live. Whether results can be distinguished — a particle found, theories separated — depends on the size of the error bar, so understanding, shrinking, and honestly reporting it is the heart of experimental physics.
- Two Kinds of Wrong: Random and SystematicDistinguish random from systematic error and the linked ideas of precision and accuracy.Random errors scatter results around the truth and hurt precision; systematic errors shift all results the same way and hurt accuracy. The dartboard shows precision and accuracy are independent, and a precise-but-inaccurate result is the dangerous case. Repetition cures random but not systematic error.
- Putting a Number on Your UncertaintyTurn repeated readings into a reported value using mean, standard deviation, and standard error, and understand the √N law.The mean is the best estimate, the standard deviation the spread of single readings (which doesn’t shrink with N), and the standard error (σ/√N) the uncertainty on the mean (which does). Because precision improves only as √N, it is expensive; results should be rounded so the value’s precision matches its uncertainty.
- How Uncertainties Travel Through Your MathExplain how uncertainties propagate through calculations and where precision is most easily lost.Absolute uncertainties add for sums/differences and relative ones for products/quotients, combining in quadrature for independent errors — so the largest term dominates and should be fixed first. Subtracting two large near-equal numbers causes catastrophic cancellation, destroying relative precision.
- The War on SystematicsPresent the hunt for systematic errors through calibration and the Millikan cautionary tale.Systematics give repeatable but wrong results and must be fought with calibration and cross-checks. Millikan’s 1913 electron charge (~1.592×10⁻¹⁹ C) was ~0.6% below the modern value (~1.602×10⁻¹⁹ C) due to a wrong air-viscosity value; the later ‘bandwagon effect’ shows the experimenter’s own bias is itself a systematic.
- Signal, Noise, and Not Fooling YourselfExplain signal-vs-noise, statistical significance, the look-elsewhere effect, and blind analysis.Averaging pulls consistent signal from cancelling noise (signal-to-noise ∝ √N). Discovery requires high significance (five sigma, fluke odds under ~1 in 3.5 million), set high partly because of the look-elsewhere effect. Blind analysis hides the answer until cuts and uncertainties are fixed, guarding against the experimenter’s own bias.
- The Instruments and Habits of the BenchSurvey the core instruments and habits of the experimental bench.Multimeters and oscilloscopes are the everyday interface between physical quantities (converted to voltages) and data; lock-in amplifiers extract tiny signals by frequency-and-phase detection; shielding and grounding manage noise before it reaches the data; and a faithful, dated lab notebook is essential for later distinguishing real effects from artefacts.
- When Getting the Error Bar Right Changed the WorldSynthesize the course through the LIGO and Higgs discoveries as triumphs of uncertainty analysis.LIGO’s 2015 detection of a spacetime ripple thousands of times smaller than a proton was believable because of detector cross-checks, systematic-noise modelling, and statistical significance; the 2012 Higgs discovery (~125 GeV) was claimed only at five sigma by two independent experiments. Both embody the course’s disciplined doubt — the creed that a measurement needs its uncertainty.
Questions this course answers
What is the ‘experimentalist’s creed’ that anchors this course?
A bare number can’t support a conclusion; the uncertainty tells you how much you might be wrong and therefore what you may legitimately claim.
Two teams both report a value of 100, but Team A says ±30 and Team B says ±2. Why are these very different results?
Whether you can separate hypotheses or claim a discovery hinges on the uncertainty, so shrinking and understanding it is the core skill.
What is the key difference between random and systematic error?
This opposite behaviour is why they need opposite cures — averaging for random error, hunting for systematic error.
Why is a ‘precise but inaccurate’ result the dangerous case?
A biased apparatus gives beautiful, consistent, wrong answers; the scatter that would warn you is absent, so the systematic must be hunted separately.
Why does taking more measurements NOT fix a systematic error?
Averaging cancels errors that go both high and low; a consistent offset survives averaging untouched, so it must be found by calibration and cross-checks.
Which quantity shrinks as you take more data, and measures how well you know the average?
The standard deviation describes the noise of a single reading and doesn’t shrink; the standard error, scaling as 1/√N, describes confidence in the mean and does.
Grounded in trusted sources
- John R. Taylor, An Introduction to Error Analysis (2nd ed., 1997)
- Philip R. Bevington & D. Keith Robinson, Data Reduction and Error Analysis for the Physical Sciences (2003)
- BIPM, Guide to the Expression of Uncertainty in Measurement (GUM, JCGM 100:2008)
- Paul Horowitz & Winfield Hill, The Art of Electronics (3rd ed., 2015)
- Richard P. Feynman, ‘Cargo Cult Science’ (Caltech commencement address, 1974)
- APS News, ‘August 1913: Robert Millikan Reports His Oil-Drop Results’ (electron charge 1.592 vs 1.602 ×10⁻¹⁹ C)
- CMS Collaboration, ‘Observation of a New Particle with a Mass of 125 GeV,’ CERN (2012); ATLAS & CMS five-sigma announcement, 4 July 2012
- B. P. Abbott et al. (LIGO Scientific Collaboration), ‘Observation of Gravitational Waves from a Binary Black Hole Merger,’ Physical Review Letters 116 (2016)
Every Wunder lesson is built from real, reputable sources — never invented.
Related Science courses
Wunder is a personalized learn-anything platform — tell it any topic and it builds a beautiful, fact-checked course in minutes, with narration, a knowledge check, and a college-style University track.
Browse more Science courses · All topics · Home
© 2026 Wunder Learning LLC · Terms & Privacy