Sampling bias

Imagine you have a bag of candy with different colors. You want to know what color is the most popular, so you take a handful out of the bag and count them. But if you only pick the candy from one side of the bag, or only the ones on top, or only the bigger pieces, you will not get a good idea of what colors are really inside the bag.

This is called sampling bias: it happens when you do not choose the sample in a random way, or when the sample is not a good representation of the whole group. It can happen by accident, when you do not pay attention, or on purpose, if you want to get a specific result.

For example, if you only ask your friends what is their favorite color, you are more likely to get a biased answer, because your friends may have similar preferences as you, or may not represent the general population. To avoid sampling bias, you need to make sure your sample is diverse, random, and large enough to give you a good estimate of the whole group.