Why an aggregated dataset?

Protecting the confidentiality of sample members is of paramount importance in the operation of the ONS Longitudinal Study. For this reason, no dataset including individual information (i.e. where each row or case represents one person) may be removed from ONS premises. However an aggregated dataset can be removed (in an aggregated dataset, each row represents one combination of characteristics, and one variable represents a count of the cases which share this particular combination).

One condition of removal is that the aggregated dataset should have no categories containing fewer than three cases. In order to produce an aggregated dataset with no case counts below three, it is necessary either to include only dichotomous variables (variables with only two values e.g. yes and no) or to include very few variables. In this dataset these two strategies have been combined. Even so, slight adjustments have been necessary to produce a dataset with no counts lower than three. This will barely affect estimates based on the data, but you should note that if the same analysis were performed in the course of an approved research project using the ONS Longitudinal Study within the ONS safe setting, slightly more accurate results would be obtained.

We hope that the experience of using this dataset may encourage you to exploit the much more extensive resources of the full ONS Longitudinal Study.

Last modified 12 February 2009