Prepare a useful set of standards
Use known concentrations spanning the intended measurement range. Dilutions must account for the final total volume. Label standards clearly and use suitable mixing. A mistake in the stock concentration affects the whole calibration rather than only one isolated reading.
Match the assay conditions
Apply the same relevant reagent amounts, timing, temperature and reading method to standards and unknowns. Use an appropriate blank for the assay background. Colorimetric methods can differ, so interpret the direction of the supplied calibration rather than assuming every glucose assay gives increasing absorbance.
Plot the relationship actually measured
Place concentration on the horizontal axis and the measurement on the vertical axis. Use repeats and a justified fit. A straight line is appropriate only where the relationship supports it; do not force every assay through zero merely because the concentration can be zero.
Interpolate before extrapolating
Read an unknown within the measured calibration range. If it lies outside that range, a justified dilution and remeasurement can provide a better estimate than extending a fitted line far beyond the standards. Keep the dilution factor recorded.
Correct the result and limit the claim
A calibration gives the concentration in the sample as measured. Multiply by a documented dilution factor to recover the original-sample estimate. The result depends on assay specificity and controls; a classroom unknown described as urine is an experimental sample, not by itself a clinical diagnosis.
Worked data example
Within a verified linear working range, standards of 2 and 4 g dm⁻³ give corrected absorbances 0.20 and 0.40. An unknown diluted fivefold gives 0.30. Estimate its original concentration.
Linear interpolation gives 3 g dm⁻³ in the diluted sample. Original concentration = 3 × 5 = 15 g dm⁻³. The fivefold factor refers to the final dilution, including the stock volume.
Why should an unknown beyond the highest calibration standard be diluted rather than simply extending the line?
The relationship has only been checked over the standard range. A documented dilution can bring the unknown into that range and reduce reliance on unsupported extrapolation.