大象传媒

Data handlingAssessing data for accuracy and reliability

It is important to make sure any information you work with is robust. Things to look out for include validity of the source, scale used, sample size, method of presentation and appropriateness of how the sample was selected.

Part of MathsData analysis

Assessing data for accuracy and reliability

A dog and a cat at a table
Image caption,
Dogs or cats: Which make the better pet?

Example situation

We want to know whether dogs or cats make the better pet so we look online and find differing views over two websites:

Two phones display websites. One is a dog website and says 'Dogs are the best pet'. The other is a cat website and says 'Cats are better pets than dogs'

Dog pals states that dogs are the best pet. CatMates says that cats are better pets than dogs.

We would need to do more research to find out if one was right and the other wrong, or if they are both wrong.

This could be an example of bias.

Bias

Bias is where a person / company / website has a particular preference. They might choose to present information to make the listener / reader see their viewpoint rather than giving a true balance of all sides.

We are probably used to hearing comments that are not robust and, hopefully, can realise when this is the case. For example:

鈥楳y aunt owns the best caf茅 in town鈥

or

鈥"We are the best football team.鈥

While these comments may, in fact, be true there is a clear bias here towards the caf茅 or football team because the people speaking have a personal connection to them.

...But if that information is displayed in a graph it must be correct, right?

Wrong!