When taking a measurement, there will always be some amount of error. The degree of error depends on the device being used to take a measurement and is referred to as its uncertainty.

Uncertainty can be discussed using two terms - accuracy and precision. They do not mean the same thing, as one might think. These two terms have been used interchangeably (and wrongly so) many, many times.

When a measurement is taken, the device used to carry out the measurement will dictate the accuracy of that measurement. If you were to measure the thickness of a coin, using a digital caliper accurate to the tenth of a millimeter would be much better than using a meter stick.

In general, the more increments (lines between lines) on a measuring device the more accurate it is. A 100.0 mL graduated cylinder has lines representing each milliliter, whereas a 100 mL beaker only has lines estimating every 10 milliliters.

beaker (left) vs. graduated cylinder (right)

In summary, accuracy can be defined as the closeness of a single measurement to is true, actual value.

Precision is another story - this term is a bit more complex to describe.

First, precision's definition - the closeness of a series of measurements (taken with the same device by the same person) to one another.

For example, if two people are measuring the volume of a glass of water using a graduated cylinder, each person may have a different technique, and a different interpretation of the results. To limit this source of error, one person must perform the measurement, and it must be done more than once in a consistent way.

Second, the more measurements one takes, the better. Since precision deals with multiple measurements, it is suggested that at least three separate measurements be taken (called triplicate measurement). The average of these measurements can then be interpreted as the best (most accurate) answer.

Source: hotdesign.com

Here are some examples of measurements and a discussion on their accuracy and precision.

A 1980 penny is weighed using two different electronic balances. Balance A reports a mass of 3.1 grams, and balance B reports a mass of 3.089 grams.
  • Since only one measurement was made, no comment can be made on the precision of the measurement.
  • Balance B provides a measurement that is accurate to the ten-thousandth of a gram, much more accurate than balance A's tenth-of-a-gram accuracy.

The 1980 penny is then weighed in triplicate using balance B, and the results are 3.089 g, 3.088 g, and 3.090 g.
  • These three measurements are all within 0.002 gram, so it is safe to say that these measurements have good precision.

What would poor precision look like?
  • Suppose a student measures 30 mL of water into a 100 mL beaker (only having ten milliliter increments), transfers the water into a 100.0 mL graduated cylinder, and records the results three times.
  • The volumes this student observes are 25.5 mL, 34.1 mL, and 29.2 mL, respectively. These measurements vary by almost 9 milliliters, a sure sign that the beaker is not a very precise device.
  • However, the average of all three measurements is 29.6 mL, which is not too bad in terms of accuracy. Now you see the importance of triplicate measurement!

In the laboratory, you will use a wide variety of measuring devices, and each will have its own degree of accuracy. Just remember a few things:
  • glassware will have volume markings etched into the glass. Read 'between the lines' - you can estimate one decimal place smaller than the smallest markings on the device. The same applies to rulers.
  • Electronic balances will provide all of the proper values of a measurement - no estimation is needed (or possible).
  • Place glassware and balances on a level surface before taking a measurement.
  • Read all liquid measurements to the bottom of the meniscus (the curve in the liquid).
  • Use consistent, repeatable techniques so as to maximize your precision.

Proper reading of a meniscus.