The self-selection bias is a type of sampling bias in which results are skewed because people are allowed to choose if they will be counted or not.
This is most clearly the case when people are allowed to seek out opportunities to participate in polls, surveys, and studies. The flip side to the self-selection bias is the non-response bias, which is simply when people opt-out rather than opt-in -- for example, when you hang up on the person conducting a phone survey.
When talking about formal and informal studies, the self-selection bias tends to exhibit itself when people with more free time (or extra-credit opportunities in undergraduate classes) participate in more studies; when you take a survey on your favorite website but not on the website you never visit; or when people with more severe forms of a disease spend extra energy trying to get into clinical trials. However, the selection bias also appears frequently in everyday life, as when your friend assures you that you will love Pokemon Go, because all the people who tried it when it came out -- Pokemon fans all -- loved it.