Socioeconomic Index for Small Areas (SEISA) - Summary

The sociologist Peter Townsend played a pivotal role in helping to develop our understanding of what we mean by deprivation and poverty. His work continues to influence thinking today (see, for example, the technical report on the Indices of Deprivation 2019 in England).

Aside from helping to identify the consequences of low income, deprivation measures do offer other advantages too.

Household income can be challenging to estimate (e.g. due to the difficulty in capturing household size and/or housing costs).

Furthermore, income data for all UK households is not available in the public domain. 

For example, during the COVID-19 pandemic, those from more deprived backgrounds may not have had the tools/equipment (e.g. laptops, reliable broadband connectivity etc) to fully engage with online learning.

This hinders their ability to reach their full educational potential.

In higher education, providers carry out various forms of widening participation activity (e.g. outreach work within the community, contextual admissions, additional study support) to promote equal opportunity for all.

The main (regular) source of information on deprivation among UK households is the Family Resources Survey. However, as this is a sample survey, this will only capture data on a fraction of the overall population and may not adequately cover all parts of the country.

It is for this reason that the Indices of Deprivation have been developed. Further details on this measure in each nation can be found via the links below.

England: Index of Multiple Deprivation

Wales: Welsh Index of Multiple Deprivation

Scotland: Scottish Index of Multiple Deprivation

Northern Ireland: Northern Ireland Multiple Deprivation Measure

Why the Indices of Deprivation are country-specific

Each country uses a different methodology in formulating its final index. For example, in Wales, eight domains are used to create the Welsh Index of Multiple Deprivation, though only seven are used in the other nations.

The weighting given to each domain in the final index can also differ between countries.

Consequently, the Indices of Deprivation are country-specific.

The potential benefits of a UK-wide deprivation measure

The lack of a UK-wide deprivation metric can lead to it being more difficult to answer questions on social mobility (as noted by the Office for Statistics Regulation and the Social Mobility Commission).

For higher education providers who recruit from across the UK, the absence of such a variable makes it harder to assess the background of individuals from different parts of the country.

Having statistics that can be compared across geographical domains can also be helpful.

For instance, it can highlight which nations/regions are meeting particular policy objectives.

From this, there is a possibility for other countries/localities to learn from their experience and apply best practice.

The questions asked in the Census are harmonised as far as possible across the UK.

Consequently, a number of the variables collected in the Census (and their corresponding outputs) are either broadly or highly comparable across all four nations. This opens up the possibility to create a measure that can be applied throughout the UK and for comparing statistics across nations (see the appendix of our technical report for further discussion on these matters).

Summary statistics on each area of the UK can be downloaded by the public.

We use the most granular geographical domain available to form our measure. These areas generally contain less than 500 individuals and are therefore smaller than those used to create the Indices of Deprivation. We refer to these as ‘small areas’ from this point onwards (though they are also commonly known as ‘output areas’).

Our literature review indicated that education and employment were key determinants of whether households are likely to have low levels of income – thus increasing their chances of experiencing deprivation.

Previous research has questioned the use of car ownership in the creation of a deprivation index, as poorer households in rural areas may have no choice but to purchase a vehicle if there is limited public transport in their area.

Our empirical work also indicated that including housing tenure did not materially improve the final measure.

Taking an average of the two proportions is justifiable in this instance as there is a strong linear correlation between the two variables (as illustrated in the chart below). Both variables also have similar distributions, which is discussed in the technical report available via the SEISA resources page.

As the comparability report highlights, the questions on qualifications are broadly comparable, whilst they are highly comparable in the case of occupation.

The highly linear correlation between the two variables we use to create our measure means that even if we only used the occupation variable to generate a measure of deprivation (i.e. we rely upon the most comparable variable), similar results/conclusions would be reached as with our current methodology.

The correlation between qualifications and occupation in small areas of the UK

Let us assume there are just 10 of these small areas from the Census.

The proportion of residents with below degree-level qualifications (Qualifications) and the proportion of residents not working in professional/managerial occupations (Occupation) is firstly calculated in each area alongside the average of these two values. As they are percentages, their value will always range between 0 and 100. A hypothetical table is provided below.

Having completed this first phase, the next step is to proceed only with the area name/code and ‘Average’ column.

These areas are then ranked based on the ‘Average’ column from largest to smallest, with areas then assigned their appropriate decile based on their ranking.

Through our research, we illustrate that having lower levels of qualification and working outside of professional/managerial roles are both factors that raise the probability of being in deprived circumstances (e.g. by being more likely to have low levels of income).

It is for this reason that higher percentages would indicate that an area is more deprived and therefore should be allocated to a lower decile.

When you select particular areas in our interactive maps or download the 'SEISA dataset' from the 'SEISA resources' webpage, you will be provided with this average percentage (referred to as the ‘SEISA proportion’) and the corresponding decile (both on a UK-wide basis and country-specific level).

Further information on why there are both UK-wide and country-specific deciles is supplied in the expanding box underneath the maps for each nation and also in the cover sheet of the ‘SEISA dataset’ excel file.

We highlight the association between our measure of deprivation and household income/housing tenure/family composition/health through a series of charts below.

Please note that we have used country-specific deciles for assessing the correlation between SEISA and income in England (‘SEISA decile England’) and Wales (‘SEISA decile Wales’). For the figures on the association between SEISA and housing tenure/family composition/health, we use UK-wide deciles (‘SEISA decile UK’), given these are based on a UK-wide analysis.

Correlation between SEISA and 2011/12 equivalised income estimates (after housing costs) in England

Correlation between SEISA and 2011/12 equivalised income estimates (after housing costs) in Wales

The association between SEISA and housing tenure

The association between SEISA and family composition

The association between SEISA and self-reported health

As can be seen from the charts below, the association is very similar to the one observed for 2011/12 income data. This suggests that the measure is still valid to use today.

Please note that we have used country-specific deciles for assessing the correlation between SEISA and income in England (‘SEISA decile England’) and Wales (‘SEISA decile Wales’).

Correlation between SEISA and 2020 equivalised income estimates (after housing costs) in England

Correlation between SEISA and 2020 equivalised income estimates (after housing costs) in Wales

In the higher education sector, we expect that the measure will be used to support widening participation activity, though we leave it to providers to choose how they wish to apply it in their work.

HESA will be looking to apply this measure in our research and official statistics outputs (e.g. when examining inequalities in the higher education sector and how these differ by nation).

As with the Indices of Deprivation in each nation, our measure may be of benefit to individuals/organisations who need data on deprivation to support their decision-making processes, as well as those who carry out research/analysis on this matter. This may include those in;

• Other parts of the education sector.
• National Government (e.g. in allocating funding towards the most deprived areas).
• Local Government (e.g. in determining where to target particular programmes/interventions).
• Healthcare (e.g. in exploring health inequalities).
• Charities/researchers/other organisations specialising in issues relating to poverty and deprivation.

Firstly, the data has been presented as part of an interactive map/postcode search tool, which can be accessed in this section of our website. You will see that beneath the maps, we provide a table of data that relates to the area you search for. This can be downloaded.

Further information on how to do so is supplied in our instructions/user guide that accompanies the maps/postcode search tool.

Secondly, separately to the maps, we have provided an excel file containing the small area (‘output area’) code, SEISA proportion and corresponding deciles (both at a UK-wide level and on a country-specific basis).

If you do not have small area (‘output area’) codes in your dataset, but do have the postcode, you can map postcodes to output areas using lookup files supplied by the Office for National Statistics. This will then provide an avenue through which to link to our own deprivation data. Please be aware of the license terms and conditions if you choose to use postcode products.

Finally, if you are asking for an extract of HESA data through the Jisc Data Analytics team, users can now request the SEISA proportion and corresponding decile (this can be the UK-wide and/or country-specific deciles depending on the nature of your analysis) as part of their request.

In this file, users will find data on the Indices of Deprivation for each nation, TUNDRA and the 2011/12 area-level income estimates for England/Wales that we discuss above, alongside the key Census variables we use in creating/validating our measure.

While users are also welcome to utilise this for their own research purposes, this is a static file and will not be updated any further (e.g. if there were to be revisions made to the Indices of Deprivation).