Infer mass through an explicit calibration model

Define the quantity before reading the display

The measurand is the quantity we intend to estimate—in this case the mass of a sample under a stated comparison. The instrument displays an indicated mass. A calibration relationship connects reference inputs to readings, and its inverse connects a reading to an estimate. Calling all three numbers “the mass” hides which quantity has been observed and which inferred.

Derive a model rather than apply a correction by habit

Use the original affine model y = s x + z. Here x is reference mass, y indicated mass, s a dimensionless scale factor, and z an additive zero offset in grams. Three exact synthetic standards at x = 0, 10 and 20 g give a straight line. The zero standard identifies z; a difference of two nonzero readings divided by the reference-mass difference identifies s. Solve algebraically: x = (y − z)/s. Subtraction must precede division.

Work an inverse example

At s = 1.05 and z = 0.10 g, the standard readings are 0.10, 10.60 and 21.10 g. The unknown display reading is 16.60 g. Subtracting the zero gives 16.50 g indicated, then dividing by 1.05 gives 15.7143 g inferred. Merely subtracting 0.10 would leave a 5% scale effect. The extra displayed digits are calculation digits, not a claim of physical measurement precision.

Make a controlled comparison

Hold z at 0.10 while changing s from 0.95 to 1.05. Predict whether the same indicated unknown corresponds to more or less true mass. Then hold s fixed and alter z. Explain each direction from the inverse formula, not from a memorized “always add correction” rule. Capture the reference table for each case and verify that the inferred unknown remains between the 0 and 20 g reference inputs.

Challenge the model

A straight line drawn through exact synthetic standards does not establish linearity of a real instrument. A withheld reference near the unknown can test an interpolation prediction. Repeating standards at later times can reveal drift. Neither check alone proves absence of all biases. If the unknown lies outside the calibrated range, extrapolation adds a different assumption; this bounded activity keeps it inside the range.

Transfer with a reversed question

A different packet has s = 0.98 and z = −0.20 g. Predict the display for a true 12.0 g sample, then invert your result. The forward prediction is 11.56 g; applying (11.56 + 0.20)/0.98 returns 12.0 g. State why a successful algebraic round trip checks consistency of the formula, not the truth of an instrument model.

Source and evidence boundary

These explanations, calculations and diagrams are original. Conceptual references checked 30 October 2026: OpenStax measurement uncertainty, OpenStax dimensional calculations. The linked textbook is a reference, not copied lesson text or imagery. All displayed datasets are synthetic teaching packets. Saved numerical agreement or prose is not a performed practical, independent reasoning review or laboratory certification. No handling, heating or chemical preparation procedure is requested.