That statement is true.
Because of this, the data that is shown by research organization may contradicts the real situation due to the incomplete samples.
For example, if this happen during political research, that campaign organizations may skip a certain type of voters segmentation and cost them the election.