Mathematics
Random Error
Quick fact
Random error is why no measurement is perfectly repeatable, but averaging many measurements reduces its impact, making the average more reliable than any single measurement.
Why this is interesting
Have you ever measured something multiple times and gotten slightly different results each time? That's random error—the natural, unpredictable variation that affects every measurement.
Read the full explanation
Understanding Random Error
Imagine shooting arrows at a target: random error is like the arrows clustering around the bullseye but with some spread due to tiny, uncontrollable differences in aim or wind. If you measure the same object repeatedly with even a precise tool, you get a scatter of values. This scatter is random error. It comes from fluctuations in the environment, the instrument, or the observer—things you can't control or predict. Unlike a mistake, random error isn't caused by a slip; it's an inherent part of measuring. The more times you measure, the more the random errors cancel out, and the average of your measurements gets closer to the true value. Random error determines the precision of your measurement—how consistent your results are—not its accuracy (which is affected by systematic error).
A deeper explanation
Random error arises from unpredictable, non-systematic factors that introduce variability in each measurement. These fluctuations are often modeled by a probability distribution, most commonly the normal (Gaussian) distribution. The spread of the distribution is measured by the standard deviation—smaller standard deviation means less random error. When you take multiple measurements, the average (mean) becomes more stable because the positive and negative random deviations tend to cancel out. The precision of that average is given by the standard error, which equals the standard deviation divided by the square root of the number of measurements. This relationship reveals a key principle: doubling the number of measurements reduces the standard error by only about 30%, so diminishing returns set in quickly. Understanding random error is critical in science and engineering—it underlies confidence intervals, hypothesis testing, and any claim about whether results are 'real' or just noise. It also reminds us that no measurement is perfect, and uncertainty must be quantified and reported.