Sociologists have several basic methods of doing research. First is the experiment, which is useful for clearly defined questions in which varying factors can be controlled. A second method sociologists use is the survey, which is useful for gaining facts about a particular group; in order to be sound, the survey must be random. Direct observation is helpful for in-depth studies of social processes, but to be useful, such observations must be made by a skilled researcher. Finally, existing information can be studied as the basis for new conclusions.