ABSTRACT
At an early stage a decision must be taken on the units of analysis, that is, the subjects on which we are going to collect information. (Once assembled in a data file, subjects are often referred to as cases.) In many studies the unit of analysis or subjects will be people (voters, employees, children, offenders, for example). Data on people is often referred to as data at the individual level. Alternatively, the unit of analysis could be an organisation, institution or region such as schools, hospitals, police forces or counties and countries. For these units of analysis, data might be collected on the characteristics of the organisation or institution, such as the type of school or hospital, whether the police force covers a rural or urban area and so on. In addition, aggregate data might be assembled for the organisation based on the individual data for that institution, such as the number or proportion of pupils passing examinations (in the case of schools) or the number of crimes per number of people living in the police force area (in the case of police forces). Such measures are thus in the form of proportions (or percentages) or rates – the latter being common when comparing countries where measures often include population density, Gross Domestic Product per capita, literacy rate and so on.
