Standard Statistical Techniques

Many of the 1998 audits were based simply on describing the numbers of incidents that happened, when and where. While this is useful, it tells nothing about relationships or significant patterns. Many previous audits omitted correlations. Of those who did, many did not go far enough in seeking to understand what was behind that correlation. So the ability to relate intervention to identified need was weakened.

For example, there is a significant correlation between Saturday nights and alcohol-related violence, and between alcohol related violence and peaks in presentation at Accident & Emergency units. This does not, however, mean that you have explained the cause but have indicated something to explore further. Some alcohol-related violence happens in the home – especially domestic violence - and other alcohol related violence happens in and around football grounds.

A correlation would suggest the need to further investigate what is going on. Binge drinking, gang rivalry, domestic violence and drug-alcohol cocktails are all factors related to alcohol-induced violence; but occur in different ways which require different responses. A good analysis of data will go some way to understanding this and help prioritise policy areas. By understanding and applying some fairly basic statistical techniques, it is possible to add value to the huge amounts of data which can be collected. This section covers:

• Writing a plan

• Levels of data

• Using software to do analysis

• Doing the analysis

• Developing an effective audit: aspiring to increased standard of audit

• Learning lessons for next time

Back to Data Analysis