1.2. Types of Errors

Systematic Errors

• Definition: Errors that occur consistently in the same direction every time you measure. They arise from flaws in calibration, bias in the measurement method, or environmental factors.
• Examples:
  • A balance that always reads 0.050 g too heavy (unzeroed balance).
  • A spectrophotometer with a wavelength offset due to misalignment.
  • A thermometer that is poorly calibrated and always reads 0.2 °C higher than the real temperature.
• Detection: Systematic errors often reveal themselves when a known standard is measured and the result is consistently off by the same amount. Comparing two independent methods can also expose systematic bias if they disagree by a constant offset.
• Correction: Recalibrate the instrument, use calibration curves with standards, or apply a known correction factor to all measurements.

Systematic errors are consistent inaccuracies in measurements. For instance, if you have a scale that always shows 0.050 g too heavy, every weight you measure will be off by this same amount—this is a systematic error. These types of errors are often caused by issues like incorrect calibration of instruments or biased measurement procedures. To spot these errors, you can measure a known standard; if your value is consistently wrong, you'll know there's a systematic issue. The good news is that systematic errors can usually be corrected by recalibrating the instrument or applying a correction factor to the measurements.

Imagine if you always weighed your bag of flour on a scale that gradually added 0.050 g to whatever the true weight is. Every time you bake, your recipes will be slightly off, potentially ruining your cookies! If you notice that your cake never rises like it should and eventually check the scale, you could recalibrate it or add a note in your recipe to subtract that extra weight, correcting the error for future baking.

Random Errors

• Definition: Errors that cause measured values to scatter randomly above and below the true value. They arise from fluctuations in the environment, electronic noise, observer reaction time, or other unpredictable factors.
• Examples:
  • Fluctuations in a digital balance reading in the last decimal place.
  • Small swings in temperature or pressure affecting a volume measurement.
  • Observer reading a meniscus with slight variability.
• Detection: Random errors manifest as spread (statistical scatter) in repeated measurements. The more repetitions, the better the estimate of the random uncertainty.
• Quantification: Use statistical measures—standard deviation, variance, standard error—to describe how much individual measurements vary.
• Reduction: Take more measurements and average them. Average has lower random uncertainty (uncertainty of the mean = standard deviation ÷ sqrt(number of measurements)). Use instruments with higher resolution; control the environment more tightly.

Random errors are unpredictable variations in measurements that can be caused by a variety of factors, such as changes in temperature or noise from electronic devices. Unlike systematic errors that are consistent, random errors cause your measurements to fluctuate in an unpredictable way, leading to a range of values around the true measurement. The more times you measure and average out those values, the smaller the effect of random error becomes. Statistically, we use concepts like standard deviation to quantify how much our measurements vary and help us express the uncertainty in our results.

Think about measuring the temperature of boiling water. If you check the temperature multiple times, you might see some fluctuation: one time it’s 99.8 °C, the next it’s 100.2 °C. These differences can be due to slight changes in the environment or just the way you place the thermometer. If you measure it 10 times and take the average, you will get a more reliable temperature that represents the actual boiling point, minimizing the influence of those small random errors.

--