15.7: Measurement Error and Accuracy

Quality and Regulatory Compliance in Bioscience · LibreTexts

0%

Contents

    Text

    Size
    100%
    Theme
    Typeface
    Line length
    Spacing

    Read aloud

    How

    Speed: 1×
    Light up
    Highlight

    Symbols: Mulberry Symbols, © 2018–2026 Steve Lee, CC BY-SA 4.0 (licence, the set), the site's own copies.

    From LibreTexts, under CC BY-SA 4.0, as the source states it; its equations are written in TeX, which the PDF shows typeset. This page at LibreTexts.

    15.7: Measurement Error and Accuracy

    Accuracy, Bias, and Measurement Error.pngMeasurement error is the difference between a measured value and a reference value. A true value is generally not known exactly, so an accepted reference value and its uncertainty often provide the practical basis for comparison. Accuracy describes closeness to the relevant true or accepted value and depends on both bias and precision.

    Gross mistakes include recording the wrong unit, using the wrong sample, or making an arithmetic error. Such events require correction and investigation. They should not be treated as routine random variation.

    Systematic error creates a consistent tendency for results to differ from the reference. Examples include a pipette that consistently delivers too little, an inappropriate calibration model, or incomplete extraction. The estimated systematic difference is often called bias.

    Random error causes unpredictable variation among repeated results. Averaging suitably independent measurements can reduce uncertainty in the mean, but it does not eliminate systematic bias. It also does not justify averaging away a failing result contrary to the approved method.

    Good technique, suitable calibration, validated methods, representative sampling, effective controls, and appropriate review all support accurate results. Calibration alone cannot correct errors caused by unsuitable sampling or sample preparation.

    One useful descriptive calculation is absolute percent error of the mean. It summarizes the difference between a replicate mean and a nonzero reference value; it is not a complete measure of analytical accuracy.

    Absolute percent error = |mean − reference value| / |reference value| × 100

    Relative bias (%) = (mean − reference value) / reference value × 100

    Relative bias retains direction for a positive reference value. A positive value indicates that the mean is higher than the reference; a negative value indicates that it is lower. Percent-based calculations are inappropriate when the reference is zero.


    This page titled 15.7: Measurement Error and Accuracy was last modified on Mon, 09 Nov 2026 00:14:02 GMT and is shared under a CC BY-SA 4.0 license and was authored, remixed, and/or curated by Emalee MacKenzie.