The overall score is made up of cleaning performance score (55%), ease of use (30%) and pet hair removal (15%).

This score is a composite of the quick carpet clean score (67%) and the complete carpet clean score (33%).

The quick carpet clean score is based on just four passes over the carpet rather than the eight performed in the deep clean test. This test simulates typical everyday cleaning. We pre-load the vacuum cleaners with artificial household dust, and embed sand evenly into a section of carpet before vacuuming it to a set pattern and speed. The sand collected is then weighed and recorded as a percentage of the sand applied; this is repeated twice for each vacuum cleaner and results averaged. We rate all models based on the dirt pick-up performance from carpet using the standard cleaning head.

The complete carpet clean score is based on eight passes over the carpet. We pre-load the vacuum cleaners with artificial household dust, and embed sand evenly into a section of carpet before vacuuming it to a set pattern and speed. The sand collected is then weighed and recorded as a percentage of the sand applied; this is repeated twice for each vacuum cleaner and results averaged. We rate all models based on the dirt pick-up performance from carpet using the standard cleaning head.

Ease of use is based on how easy it is to manoeuvre the cleaner, and move the cleaning head on the carpet and under low furniture. We assess how difficult it is to empty and change the dirt receptacle; operation of the controls; and how easy the supplied accessories are to use and store.

To test this, sand is spread into a right-angled corner. Using the standard cleaning head, it's vacuumed to see how close to the corner each cleaner is able to pick up dirt.

To test pet hair removal, we embed cat hair in the carpet and, as with the dirt pick-up test, vacuum it to a set pattern and speed. We rate the results based on how much cat hair is left behind after the test. Where supplied, the turbo brush or power head is used for this test; otherwise, we use the standard cleaning head. Pet hair for testing supplied by Woof Cat & Dog Grooming, Bondi, NSW.

Noise is measured at the user's position at maximum power setting. For comparison, a typical conversation is carried out at about 60dB, while city traffic is measured at around 80dB.

Scores are taken from our latest brand reliability survey. It does not add to the overall score. The score refers to the brand, not a particular model. NA or Not available means there weren't enough responses for that brand to get a score.

Scores are taken from our latest brand reliability survey. When rating satisfaction we ask owners to tell us what they thought of the products they own. We use this information to give each brand an owner satisfaction rating so you can see what people who use the brand day in day out think of them. NA or Not available means there weren't enough responses for that brand to get a score.

We recommend barrel and upright vacuum cleaners that score at least 75% overall, and at least 70% for both quick carpet cleaning and complete carpet cleaning.

Recommended or typical retail price. You can often get a better price by shopping around.

For bagged models, the average cost per dust bag.

The vacuum cleaner's stated power consumption (usually its claimed maximum power).

The power consumption we measure in the lab, on the maximum power setting after a two-minute running-in period. This can be different from the stated power consumption - this may be due to differences in how manufacturers measure it. Take the stated power consumption as a guide only. 

The bag or bin's volume in litres, measured in our lab.

This figure incorporates the distance of the suction hose and unit, as well as the power cord.

What we received out of the box. You may be able to order additional accessories/attachments at extra cost.

The model has a claimed origin. This ought to be taken with a grain of salt, given it's rare that all the materials for a product came from the one country.

The standard head for many vacuum cleaners, this head can be used on either carpet or hard floor, and usually has a foot-operated switch to adjust between the two modes. It usually has brush edges that lower in hard floor mode to protect the floor.

Has a motorised brush and is a good option for quickly cleaning up pet hair and lint as well as extracting dirt from carpet.

Has a rotating brush driven by the vacuum's air flow. Good for picking up pet hair and lint, though generally less effective than a power head.

A head or tool attachment specifically for cleaning hard floors.

A long, narrow attachment for cleaning narrow gaps (e.g. between sofa cushions).

An attachment for cleaning soft furnishings such as curtains, sofa cushions and so on.

An attachment for dusting and vacuuming hard furniture, along skirting boards and architraves, and so on.

It's more convenient to store the tools and attachments in or on the body of the vacuum cleaner than to have to keep them separately elsewhere. Some models only have storage for one or two of the supplied tools.

A small hole with a sliding cover on the wand that allows you to reduce the suction for delicate cleaning jobs by letting some air escape through the hole.

A control on the body or handle of the vacuum cleaner that lets you set different power levels for different jobs, from delicate cleaning of upholstery to more powerful suction for carpet or hard floors.

High-Efficiency Particulate Air filtration - a filter for trapping very fine dust and pollen. Useful if you have allergies or asthma. Some are washable, others need to be replaced every year or so.

A washable HEPA filter means you won't need to buy a replacement, but remember to clean it regularly for maximum performance.

An indicator that the bag is full and needs replacement (or a clear bin that lets you see when emptying is needed).

The vacuum can switch from suction to blowing out air. Handy for inflating things like airbeds and balloons, or for blasting dust from hard-to-reach areas.