The paper offers the overview of sampling procedures used in studies on divorce, limited to people who get divorced with extension to pre- and post divorce periods. It presents various types of random and non-random sampling, their advantages and drawbacks in studies on divorce. In a relatively broad overview of non-random procedures an attention is drawn to the fact that non-random sampling, often used in the case of no other alternative, usually does not allow for an statistical inference and has only a limited value for generalization of conclusions. In the final part sources of data used for the secondary analyses are discussed at length. Numerous examples of surveys are presented, which have been designed for different purposes, and are used for studies on divorce in various countries, like for instance the General Social Survey and the Panel Study of Income Dynamics in USA, the General Social Survey in Poland. That type of data source is becoming more and more important.