Random sampling requires an institution to use only a limited sample. As its name implies, the source from which data are to be drawn is identified randomly using a mathematical theory of probability. It typically requires the use of statisticians in survey design and interpretation. For instance, there are 100 communities in a region and civic education sessions on electoral rights are supposed to be held in each. The institution wants to monitor this to ensure that the education is in fact carried out, and that it is fair and complete. Rather than monitor all civic education sessions in each community, an institution could randomly select 10 communities—for example, by drawing names of communities out of a hat—that it will monitor and draw inferences from those results.