Measurement (random) error

All measurements are subject to error, but the dangers of misinterpretation are reduced by recognizing and understanding the likely sources of error and by adopting appropriate protocols and calculation procedures.

A common source of random error is carelessness, e.g. reading a scale in the wrong direction or parallax errors. This can be reduced greatly by careful recording and may be detected by repeating the measurement. Other errors arise from faulty or inaccurate equipment, but even a perfectly functioning machine has distinct limits to the accuracy and precision of its measurements. These limits are often quoted in manufacturers' specifications and are applicable when an instrument is new; however, you should allow for some deterioration with age.

One major influence virtually impossible to eliminate is the effect of the investigation itself: even putting a thermometer in a liquid may change the temperature of the liquid. The very act of measurement may give rise to a confounding variable. You should include descriptions of the possible sources of error(s) and estimates of their importance in any report and these should not be used as an excuse for poor technique or inadequacies in your experimental design.