No experimental measurement can possibly be perfectly precise. Take, for example, a wooden stick that is approximately two meters long. If a scientist were to measure that stick with a ruler marked only with meters, then he could only conclude with certainty that the stick measured 1 meter (though of course he would recognize that his measurement was inexact). If his ruler was marked with decimeters, then he could see with certainty that the stick measured 1.1 meters. If he could measure centimeters, he might see that the stick actually measured 1.12 meters. Using a ruler with millimeters he could see the stick is actually 1.121 meters long. Each smaller measurement allows the scientist to determine the length of the stick with a bit more accuracy. But no scientist can use a ruler to great effect for distances much smaller than a millimeter; such small distances are simply beyond the ability of the scientist's ability to see. At some point his measurements will necessarily become slightly inaccurate.
Scientists account for this unavoidable uncertainty in measurement through the use of significant digits. Significant digits do not remove the uncertainty; instead they alert others as to where the uncertainty lies. In the case of our measurement of the stick, the value 1.121 meters alerts the next scientist to come along that the last 1 digit on the right might be slightly inaccurate.
Five rules govern significant figures:
These rules ensure accurate representation and interpretation of data. If, for example, you were to read of an experimental reaction in which the resulting chemical weighed 0.0254 g, you would know that the measurement is accurate to 0.0001 g and contains 3 significant figures.
When making calculations, significant figures become very important. You must always be careful to remember how many significant figures your separate values have. The rules governing addition and subtraction, and those governing multiplication and division are a little different.
Addition and subtraction of significant figures follows a simple rule:
The final value must have only as many decimals as the original value with the least number of decimal places.