So-called accidental sampling. Please see the link.
Chat with our AI personalities
The related web sites give a good idea of the types of non-random sampling. These include snowball, convenience, quota, self-selection, diversity, expert, and others. Non-randon sampling is usually done because it is less expensive, easier, and quicker than random sampling.
Advantage -- Less effort, cost, work Disadvantage -- Less accuracy, information, difficulty of establishing true 'randomness" in some samplings.
You get a non-random sample and any analysis based on the assumption of randomly distributed variables is no longer valid. In particular, your estimates of any variables are likely to be biased and your error estimates (standard errors or sample variances) will be incorrect. Any inferences based on statistical tests will be less reliable and may be wrong.
Not less than double the highest frequency component of the signal you're sampling.
A random sample is better than a census because it takes less time and costs less.