NPL has developed six guiding principles which you should follow if you want good measurement results:

1. Make the right measurements
2. Use the right measurement tools
3. Use the right measurement procedures
4. Use the right people to do the measurement
5. Review your measurement regularly
6. Demonstrate how consistent the measurement is

Remember - no measurement is, or ever can be, exactly right.

But that's OK because none of us ever needs measurements to be exactly right; it is enough that they have whatever accuracy we need, or uncertainty we can tolerate. It is important to understand the uncertainty associated with a measurement, and that it is at a level appropriate to the application.

Make the right measurements

A measurement is made for a reason, and that reason needs to be clearly defined and understood if the measurement is to be a good one.

This is of course especially important when the measurement is being carried out for someone else.

When a system which involves routinely repeated measurements is being devised (such as a system to measure the sizes of fruit and vegetables), a pilot study is a useful first approach, to discover where any problems lie and where improvements can be made to the procedure.

Use the right measurement tools

The measuring instruments you use need to be appropriate for the task, in a good state of repair, and calibrated - and they need to be used according to the instructions of their owner or manufacturer.

Use the right measurement procedures

As there are so many factors which need to be addressed to ensure that the result of a measurement is a good one, it's vital that important or complex measurements are carried out in accordance with written procedures.

Though these might simply be the documents supplied by the manufacturer, these may not be sufficient, especially where a number of different pieces of equipment are involved.

An important function of a written procedure is to safeguard health and safety, so it will often be backed up by a risk assessment for this purpose.

You must regularly review your measurement procedures to ensure they remain appropriate.

Use the right people to do the measurement

Human error is a major source of poor measurement quality. Whoever makes the measurement (sometimes referred to as the 'operator') needs to have received the right instructions and training.

For complex measurements, this training will often include formal qualifications.

Where a group of operators is involved, their individual roles and responsibilities need to be formally agreed and clearly understood.

Review your measurement regularly

Measuring instruments are often easily damaged and their performance frequently changes as time passes, so they need to be checked.

These checks should be carried out at regular intervals rather than just before they are needed, to avoid delays.

Since many individual instruments may be involved in making a measurement (to check environmental conditions, for instance), a written schedule is usually essential.

In many cases this schedule will include both internal checks and less frequent external assessments.

Demonstrate how consistent the measurement is

A measurement result isn't much use if it's only valid at the place where the measurement is made.

For highly accurate measurements, many local factors may need to be taken into account if this is to be avoided.

The force of gravity, for example, varies by up to 1% from place to place on the Earth's surface (and changes with time too, as a result of the shifting gravitational influences of the Sun and Moon, among other things).

In turn, this affects the weights of objects (as measured, for example, by spring balances).

So, an assessment of such factors should be made in planning an important measurement.

What happens next depends on the outcome of the assessment - it may be that the factors are too small to significantly affect the uncertainty of the results.

Or they can be corrected for by the user, given appropriate data and instructions.

Or it may be that the result should be quoted with higher uncertainty.

In addition to this approach, for important or difficult measurements, other operators in other laboratories should carry out the same measurements and their results compared.

Depending how different those results prove to be, it might again be necessary to make corrections, increase quoted uncertainties, or carry out further investigations.

Search NPL's database of over 100 specialised measurement guides

Join NPL's Measurement Network

Attend NPL measurement training events