Two families of methods
If you stand back and survey every method ever invented, they fall into two grand families. The older family is wet chemistry: you measure your quantity using reactions in solution and simple, classic tools — a balance, glassware, a colour change you watch with your eyes. Weighing a solid you have made the analyte turn into, or dripping in just enough reagent until a colour flips, are wet-chemical at heart. These methods are old, but far from obsolete; they are often beautifully direct and need no electricity.
The younger family is instrumental analysis: you let a machine measure a physical property of your analyte — how it absorbs light, conducts electricity, or bends in a magnetic field — and read the signal off a dial or screen. Instruments shine where wet methods strain: they can detect fantastically tiny amounts (essential for trace analysis), run fast, and handle hundreds of samples. The catch is that they cost money, need calibration and upkeep, and can hide their assumptions behind a clean-looking number.
A shared scorecard: figures of merit
With so many methods across both families, how do you compare them fairly, or decide if one is good enough for your task? Analysts use a shared scorecard — a set of numbers that summarise how a method performs. These performance numbers are the figures of merit. Think of them like the spec sheet for a camera: not the photos themselves, but the honest numbers (resolution, low-light ability, speed) that tell you what the camera can and cannot do before you buy it.
You do not need to memorise the list yet — whole later rungs are devoted to each one — but it helps to recognise the cast of characters now, because they recur on every page from here on.
- Accuracy — how close the answer is to the true value (does it hit the bullseye?).
- Precision — how closely repeated measurements agree with each other (are the shots tightly grouped?).
- Sensitivity — how much the signal changes when the amount of analyte changes a little.
- Detection limit — the smallest amount the method can reliably tell apart from nothing.
- Selectivity and range — how well it ignores interferents, and over what span of amounts it works.
Accuracy and precision are not the same thing
Of all the figures of merit, the pair beginners most often blur is accuracy and precision — so fix the distinction now with a dartboard. *Precision* is how tightly your darts cluster together, whether or not they are near the bullseye. *Accuracy* is how close they land to the bullseye itself. You can be precise but wrong: three darts in a tight little cluster up in the top-left corner — beautifully repeatable, and beautifully far from the target. You can also be roughly accurate but sloppy: darts scattered all around the bullseye, averaging out near it but individually unreliable.
Why insist on the difference? Because they fail for different reasons and call for different fixes. Poor precision comes from random jitter — small uncontrolled wobbles you can shrink by measuring more carefully or averaging more readings. Poor accuracy that hides behind good precision comes from a consistent bias — a miscalibrated balance, a contaminated reagent — that no amount of repeating will reveal. The most dangerous result in the whole field is one that is precise but biased: it looks confident and is quietly, consistently wrong.
The map you carry from here
You now hold the whole map of this domain in miniature. There is *stuff* (a sample), a *target* in it (the analyte) surrounded by *everything else* (the matrix). You answer *which* (qualitative) and *how much* (quantitative) by coaxing out a *signal*, following a careful *procedure*, using a method from one of two great families — wet or instrumental — chosen for good *aim* (selectivity) and judged by a shared scorecard (figures of merit). Every later rung simply zooms into one corner of this map. Keep this picture, and you will never feel lost.