A number is a claim, not a fact
Imagine you weigh a pinch of salt and the scale reads 1.0 grams. It feels like a hard fact. But press a little harder: is it *exactly* one gram, or somewhere between 0.95 and 1.05? Would a better scale read 1.013? Every act of laboratory measurement is really a claim — 'as far as my instrument can tell, the value is about this.' Learning chemistry partly means learning to read that claim honestly instead of treating numbers as if they fell from the sky.
Why does this matter so much in physical chemistry? Because the whole field is built on comparing numbers — a predicted boiling point against a measured one, the energy a reaction releases against the energy a theory says it should. If you do not know how trustworthy each number is, you cannot tell a real disagreement from harmless wobble. The skill of saying *how sure you are* is just as important as the value itself.
Uncertainty: the wobble around the number
If you weigh the same object ten times, you will not get the same answer every time — 1.02, 0.99, 1.01, and so on. That spread is real, and we name it: the measurement uncertainty is the honest range within which the true value probably sits. We usually write a result as *value ± uncertainty*, like 1.01 ± 0.02 g. The little ± is not an admission of failure; it is the most scientifically mature part of the whole statement.
Uncertainty comes from two kinds of source. *Random* error is the jitter that pushes you a little high one time, a little low the next — averaging many readings tames it. *Systematic* error is a steady lean in one direction, like a scale that always reads 0.05 g too heavy; averaging never removes it, and only careful checking can. Good scientists chase both, but they fear the systematic kind more, because it hides in plain sight.
Significant figures: don't write digits you can't back up
Here is a temptation a calculator hands you constantly. You divide 1.0 g by 3 and the screen proudly shows 0.33333333. But your original mass was only good to two digits, so reporting eight is a fib — it pretends to a precision you never had. The honest digits, the ones your measurement actually supports, are called significant figures. The rule of thumb: a result can be no more precise than the *least* precise number that went into it.
- Count the trusted digits in each measured number you start with (1.0 g has two; 1.013 g has four).
- Do the full calculation with all the digits — don't round in the middle, or small errors pile up.
- At the very end, round the answer to match the least precise input (for multiply/divide, the fewest significant figures).
Calibration: pointing your ruler at a known truth
How do you catch a scale that reads 0.05 g too heavy? You weigh something whose mass you *already know* — a certified reference mass — and see if the instrument agrees. Adjusting an instrument so that its readings line up with known standards is called calibration. It is the antidote to systematic error. Every serious instrument, from a kitchen thermometer to a million-dollar spectrometer, is only as trustworthy as its last calibration.
Thermometry — the careful measurement of temperature — is a lovely example. We anchor thermometers to fixed points nature gives us for free: pure ice water sits at 0 °C, water boils at 100 °C at standard pressure. A thermometer that reads those two anchors correctly can be trusted in between. Without those anchors, a thermometer is just a tube of liquid making unverifiable claims.
Putting the habits together
Picture a tidy little experiment: you measure the temperature at which a sample melts. First you *calibrate* your thermometer against ice and boiling water. Then you take several readings and watch their *scatter* to estimate your uncertainty. Finally you report the melting point with the right number of significant figures and a ± range — say 54.3 ± 0.2 °C. That single line carries a value, an honesty about precision, and a promise that you checked your ruler. That is what a measurement should look like.