Audit Commission

Skip to content Go to accessibility page

Summarising data

Summarising data can help you make the most of numbers by spotting and understanding patterns and pictures.

Essential facts about summarising data

  1. The average is an important reference point. How something compares to the average, is usually the first thing to think about.
  2. Different ways of measuring the average end up with different results and you need to think about this before drawing conclusions.
  3. The mean average is the most often used type of average - but for some types of data the median, mode or frequencies might work better.
  4. Knowing about the average, the dispersion and polarity of the data can help draw conclusions about how well an authority is doing.
  5. The way a set of data is distributed can help you understand the make up of the data and where a specific value lies within the data.
  6. Often you can draw a conclusion about a data value by finding out which quartile it lies in.
  7. Box-plots and histograms are a useful way of quickly illuminating how a dataset is distributed. For example, if it has a normal distribution, or is skewed.
  8. Using the standard deviation will allow you to look at the distribution of the data, - and also identify outliers and unusual values.
  9. Sometimes there are too many data values to use - so you need to sample. If you do you will need to know about confidence intervals.
 
 
9 April 2009
Often when you first look at a data item, your initial point of comparison is where it is in relation to the average.
16 April 2009
This section looks at how to use frequencies and grouped frequencies (groups).
16 April 2009
Data can take many forms. You therefore need to know a few things about the data you are using before you can make a sensible judgment about how good or bad something is.
16 April 2009
When you look at a performance indicator, you often want to know how it compares with other results.
15 April 2009
Before you draw a conclusion about a performance indicator, you often want to look at how it compares with other results in the data set.
15 April 2009
It is not always possible to measure or determine the value of every possible case - i.e. the entire population.