Skip to main content

You are here:

Case Study: An Introduction to Indicators on the Neighbourhood Statistics Website

Main Content

An Introduction to Indicators on the Neighbourhood Statistics Website

1. Introduction

This case study introduces the concept of indicators, and outlines the reasons we may wish to create them. We will explain what the term means in relation to Neighbourhood Statistics and will look at examples of the different types of indicator data available on the website.

The information provides a foundation for more detailed studies, which show how to create specific types of indicators, and consider issues related to use and interpretation. Links to these can be found within this document.

This study will be of interest to anyone wishing to understand more about indicators and is available as a pdf document (246kb)

2. What is an indicator?

Definition

A statistical measure which summarises information about a particular item or subject of interest, e.g. crime or education, and which provides a more accurate and reliable indication of change in that item than the use of raw data on their own.

Background Information

Many of the datasets found on the Neighbourhood Statistics website are counts of people, properties, events or incidents, (for example, benefit claimants, commercial premises, or crimes in an area). Whilst publishing these counts is extremely useful, to obtain maximum value from the data we must also think about the reasons people want information, and how it is going to be used.

The Neighbourhood Statistics Service publishes data so that people can find out more about the social and economic conditions in local areas. The data are also used to help make decisions on the targeting and allocation of resources to areas of greatest need.

In order to make these decisions, policy makers and neighbourhood renewal practitioners need to know how conditions in areas compare to others, and how they are changing in response to measures that have already been put into place. However, counts are often unreliable on their own as a basis for decision making, because they do not take into account other important information about the area (for example the size and structure of the area's population, the density of its housing, or its physical geography).

In contrast, a good indicator gives us the ability to compare an area with others, or to compare change over time. Indicators can tell us how an area is 'performing' in relation to other places or the national average, and whether it is improving or getting worse. They can also summarise complicated information so that it can be used more easily for monitoring purposes.

It is important to point out that indicators do not tell us everything we need to know about an area. They can sometimes oversimplify the situation in relation to a concept that could be quite complicated and requires detailed knowledge about a range of factors. Therefore it is generally good practice to publish underlying count data alongside indicators, so that people have access to more detail if needed. Thorough metadata (descriptive documentation) should also be provided, explaining how to interpret the statistics and any limitations in what the data show.

Indicator Types

It should be noted that the term 'indicator' is subject to some ambiguity, and often depends on the issue that is being investigated. However, the types of data that are most commonly considered to be indicators include:

In some cases though, count data may actually serve as a useful indicator, for example where the frequency of events is generally rare but in some areas may be much more common (e.g. certain types of crime). In this situation the count data might quickly indicate 'hotspots' that can be targeted for further investigation.

Finding Indicators on Neighbourhood Statistics

There are different ways to locate indicators on the website:

Back to top

3. Simple 'summary' Indicators

Summary Percentages

The most basic type of indicator generally involves deriving a proportion in relation to the total figure, and is often referred to as a 'comparative' or 'summary' statistic. They are generally composed of two main components,the numerator and the denominator.

The denominator can be thought of as the total group that could potentially have the characteristics we are measuring. The numerator is the group which we know actually have those characteristics.

The formula for calculating a summary proportion is (numerator/denominator). Summary proportions are often converted to percentages by multiplying by 100.

Summary indicators are usually found alongside count data in Neighbourhood Statistics datasets, as both the numerator and the denominator are available from the counts. For example, if we look at Working Age Client Group (WACG) data it can be seen that additional percentage columns based on the total number of claimants are included with the counts.

Working Age Client Group provides a count of the working age population who are in receipt of state benefits, broken down by the primary type of benefit claimed, sex and broad age group. The data are collected by the Department for Work and Pensions. The WACG data on Neighbourhood Statistics contains percentages of people claiming different types of benefit, percentages who are male and female, and percentages by broad age band.

Table 1 shows a small extract from WACG (August 2003) for three Local Authorities. This shows the total number of claimants of all benefits, the number of male and female claimants, and the percentage of male and female claimants in relation to the total.

Table 1: Counts and Percentages of Key Benefit Claimants by Sex, August 2003 (Source: DWP Working Age Client Group data)

Table 1 : Counts and percentages of key benefit claimants by sex, August 2003

Back to top

It can be seen that by calculating percentages we can easily discover the balance of male and female claimants for each area. We can also compare the percentages between areas, because they take into consideration the total number of claimants. Further, if we wanted to see how the balance of male and female claimants has changed over time, we could show the percentage increase or decrease within a given period.

This simple example shows how we can use basic summary statistics to add value to count data, and also serves to highlight the importance of including the underlying counts. In the example above, if we did not have the count data we would not know that the overall number of claimants was significantly higher in Southampton than in Durham or Gloucester. The percentages would not reveal this and would not provide a reliable basis for provision of resource and care. For example, the counts will affect the total amount needed to be paid in benefits, and the amount of resource required to process claims and pay clients.

4. Simple indicators derived from multiple sources.

Background

If we wanted to investigate a wider issues such as 'prevalence of benefit claimants' in an area, the counts and the summary statistics shown previously may not provide adequate information. For example, we have seen that there are many more benefits claimants in Southampton, but we do not know enough about the other characteristics of the area to say whether this is comparatively high or low.

Percentages Derived from Multiple Sources

Therefore, we will now look at creating indicators which take into account other information about the area. To demonstrate, we will again use Working Age Client Group data, but instead of using the total number of claimants as our denominator, we will include information about the population of the area to provide context to the counts.

In the previous example, our choice of denominator was simple as it was located within the dataset and was part of the same data collection exercise. When we are looking for a denominator from another source though, we must be much more careful about our selection. We must make sure that an appropriate source is chosen, which will provide meaningful and accurate information. More detail about selecting appropriate denominators can be found in the related case study: Using Neighbourhood Statistics Data to Create Simple Indicators from Multiple Sources'.

In this instance we wish to ascertain the prevalence of benefit claimants in relation to the number of people living in the area. Further, the WACG data relate to people of working age only (males aged 16-64 and females aged 16-59), so we need to make sure we have the working age population only in our denominator.

This information is available via population estimates data produced by the Office for National Statistics (ONS), which provide counts of the resident population by age and sex. A variety of population estimates are available on Neighbourhood Statistics, and the data presented here have been aggregated to Local Authority level from: 'Resident Population Estimates by Broad Age Band for Middle and Lower Layer Super Output Areas, Mid 2003'

Table 2 shows the working age population of each area, the total number of benefit claimants, and the percentage of benefit claimants in relation to the population.

Table 2: Benefit Claimants in Relation to Working Age Population, August 2003 (Source: DWP Working Age Client Group data)

Table 2: Benefit claimants in relation to working age population, August 2003

We can see that although the total number of claimants is much higher in Southampton, this does not necessarily indicate a greater prevalence of benefit claimants. In fact the percentage of claimants in relation to the population is the same in all three areas.

Indicators and Maps

Presenting the data as percentages is especially helpful where we want to show the figures on a thematic map. If we just used the counts we would be likely to highlight places where populations are highest, since the number of benefit claimants in an area is broadly related to the size of the population. By mapping the percentages though, we can see the prevalence of benefit claimants in relation to population size.

Figure 1 shows the difference between presenting counts of claimants (left), and percentages of claimants in relation to the population (right).

Figure 1: Maps Showing Counts (left) and Percentages (right) of Key Benefit Claimants, August 2003 for Local Authorities in London (Source: DWP Working Age Client Group data)

Figure 1: Indicator map showing counts Figure 1: Indicator map showing percentages

It can be seen that although the maps show some similar general patterns, the detail between areas can be quite different. For this reason it is recommended that percentage data are used instead of counts if thematic maps are being produced, particularly if using choropleth mapping. More information about using maps for analysis can be found in the related Case Study: Using maps to show geographical patterns in health data.

Rates

In some situations rates are shown rather than counts or percentages because they take into consideration the population at risk. For example, in the 'Teenage Conceptions' data available on Neighbourhood Statistics, the data are shown as rates per 1000 women aged 15-17. The denominator is the 'at risk' population (the number of 15-17 year old women resident in each area, based on ONS population estimates data), and the numerator is the number of conceptions to girls aged under 18.

Teenage conceptions data are based on counts and rates of conceptions to women aged under 18 who are usually resident in England and Wales. Conceptions statistics are estimated using birth registrations and abortion notifications. Those pregnancies which lead to spontaneous abortions (ie, miscarriages) or illegal abortions are not included in the estimates.

The Teenage Conceptions data are a good example of how indicators can measure progress towards objectives. The data are used to monitor the government target to reduce the teenage conception rate by 50% by 2010 (using 1998 as a base year). The rate is monitored as opposed to the actual number of conceptions, because the overall size of the population will vary between years. The rate can also be used to compare differences between areas and in relation to the national average.

Back to top

5. Using Indicators to Measure Broader Issues

For renewal practitioners and other decision makers, it is often difficult to measure the desired issue directly, and thought must be given to choosing the most appropriate indicator. For example, "the percentage of benefit claimants may offer a useful proxy measure to help assess work deprivation in an area, but it is too narrow to fully measure work deprivation on its own. There are a variety of other factors that are important such as the number and type of employers in the area, the skills and education of the workforce, and the distance and ease of travel to work.

In practice, an indicator will seldom measure a complex topic completely, and in most cases the best available proxy data source(s) must be chosen. The relevance of the indicator to the subject being measured is an important consideration, and is discussed in more detail in the related case study: Using Neighbourhood Statistics Data to Create Simple Indictators from Multiple Sources'.

Combining Indicators from Multiple Sources

It is sometimes possible to combine a number of specific indicators together so that more general issues can be measured. The most significant example on Neighbourhood Statistics are the English Indices of Deprivation 2004 data, produced by the Department for Communities and Local Government (DCLG).

The Indices of Deprivation 2004 consist of ranks and scores for 'Lower Layer Super Output Areas' and 'Local Authorities' in England, each related to a particular 'deprivation domain' (e.g., income, employment, education, health). Underlying indicators are combined to produce the measures for each deprivation domain, and the domains are combined to provide an overall 'Index of Multiple Deprivation' (IMD) rank and score. The IMD indicates overall how deprived each area is on a scale, and is a powerful tool that can be used to compare areas and allocate resources.

The methodology used to create the Indices of Deprivation is quite complex, and is detailed in the documentation associated with the data (accessible using the links above). The underlying indicators are also published via Neighbourhood Statistics, and can be used alongside the Indices themselves to provide more detailed information on a particular subject. For example, if an area ranks among the most deprived for a particular domain, underlying indicators can be used to understand the problem in more detail.

Using the Neighbourhood Statistics website, people can type a postcode into the 'Neighbourhood Summary' section on the home page, and a variety of statistics are shown about the relevant Lower Layer Super Output Area (LSOA). On the front page of the Summary, the relative places on the Indices of Deprivation are shown graphically as follows. Figure 2 shows the overall IMD rank for the LSOA Fareham 011E, and the ranks for the Income, Employment, Health, and Education domains.

More detail about the Indices of Deprivation are provided via a specific page on the Neighbourhood Statistics website.

Figure 2: Screen Capture Showing Indices of Deprivation Ranks for LSOA Fareham 011E, 2004 (Source: Indices of Deprivation Data)

Indices of Deprivation image

6. Conclusion

We have seen examples of the types of indicators available on the website, and have explored some of the reasons we use indicators instead of count data. There are also a number of issues that are not covered in this introductory document, which should be considered if indicators are being created. For example, the impact of using rounded data, the use of numerators and denominators from different time periods, and (importantly) whether the indicator actually measures the intended topic accurately. These and other issues are outlined in more detail in the related case study: 'Using Neighbourhood Statistics Data to Create Simple Indicators from Multiple Sources'.

Back to top