Data where the sampling units fall into natural groupings and these groupings may, in turn, fall into further natural groupings. For example in a survey of school students, students in a particular school are a natural grouping, and the schools in a particular school district form another level of the hierarchy.
Although most hierarchical data is nested, this is not always the case. For example, school students can be viewed as being naturally grouped by school, but could also be naturally grouped by year or by socio-economic background.
The most common type of hierarchical data is a panel, which has two levels (e.g., households and people within the household).