Representativity is to ensure that the results of a study apply to the total population. Only when it is tested a statement can be made about this aspect of research.
In every study it is almost impossible to collect data from the total population. Although this is perhaps the best way, it is not necessary. When you collect information from a limited number of people, it seems reasonable to draw conclusions for everyone in the population.
In science it is not allowed to make statements that are not substantiated or proven. So how can you be sure that the conclusion to be drawn based on the information from the respondents is valid for the total population? This is the issue of representativeness.
An example of representativity
A number of steps usually have to be taken to draw a sample. The population must first be defined. A definition of, for example, the population of London can be all residents. However, some people have a home in London, but do not live there. Others do not rent or own a (legal) home, but live there. So this definition is a bit trivial. That triviality is reduced by changing the definition to 'all London residents who are legally registered'. A sample is taken from this register. This can be done in various ways. The most important ways can be found in our paper Nine ways to draw a sample. All people in the sample are approached to participate in the study. The people who participate are the respondents. They provide the information needed to draw conclusions for the total population. This response must therefore be representative of the total population, i.e. all residents in London.
However, in each of these steps, something can go wrong that can lead to a non-representative response. Perhaps there was no response from the Soho district, so the information from the residents of these districts is missing. It is therefore impossible to draw a conclusion for all London residents. Or for example, in the West End the response was 10 times as high as in the other districts. Now the conclusions to be drawn mainly reflect the opinions of the people in the West End than of the other districts. You should therefore always check whether the response group corresponds to the total population. This must be done by testing it. The way to do this is explained in the paper Representativeness.
Finally, keep this in mind: the response must represent the population. This really needs to be emphasized, as many authors write that the sample must represent the population. However, it is easy to make the sample representative, but it is difficult to make the response reflect the population. You take your own sample and can handle it correctly. It is much harder to get people to work together. That is why it is much harder to realize a representative sample.