While collecting samples for a survey, it may happen that the sample statistic over or under estimate a population parameter. It can be called a biased sample as it is not the true representation of the population.
The bias may occur due to error in the sample collection method. It may be possible that a particular group of respondents are under represented or over represented in the sample or the individuals chosen for survey are unable to participate in the survey or the sample is too small. The biased sample can lead to fallacious results and biased estimates.
Example: If a survey to identify the most popular sports in a country is done and the respondents are chosen as viewers in a cricket match stadium, it is very likely that the sample would be biased and will produce erroneous results.