Written submission from the Office for National Statistics (GRT0066)

1. Executive Summary

The Office for National Statistics (ONS) is the UK’s National Statistical Institute, and largest producer of official statistics.  We aim to provide a firm evidence base to inform decision-making and support democratic debate.

ONS collect and publish a range of data on ethnicity.  There is a strong demand for these data, from central and devolved government, expert, community and special interest groups, local and regional government, and local service providers.  Statistics on ethnicity are used for a wide range of purposes, including to inform resource-allocation and service provision, to enable public bodies to meet their statutory obligations, to represent the interests of specific groups, and for academic research.

This short note summarises what official statistics are currently available on the Gypsy, Roma and Traveller (GRT)[1] communities specifically.  In doing so it:

• explains what information is available on GRT communities from the 2011 Census, survey sources, and administrative sources;
   
• explores the challenges of collecting data on GRT communities, and the limitations of existing data sources; and
   
• sets out work underway across the Government Statistical Service relating to the collection of data on GRT communities.

2. Official Statistics on Gypsy, Roma and Traveller communities

Currently, the 2011 Census is the primary source of statistical information on GRT communities.  This section explains why and how ONS decided to capture information about Gypsy and Irish Traveller communities within the 2011 Census, and summarises what Census data are able to tell us about GRT communities. It also provides a brief overview of other surveys conducted by ONS which capture data on ethnicity (including the Gypsy and Irish Traveller communities).
 

2.1 The 2011 Census

Background to the inclusion of a question in the 2011 Census
 

For the first time, the 2011 Census ethnic group question included a dedicated tick box for the ethnic group Gypsy or Irish Traveller. The inclusion of a dedicated tick-box for the Gypsy or Irish Traveller communities within the 2011 Census followed extensive consultation, which highlighted a strong requirement for this information: 

• In preparation for the 2011 Census ONS held a formal consultation on census content, which began in May 2005.  During the consultation, respondents demonstrated a strong requirement for data on ethnicity and several submissions expressed a specific need for a question on the Gypsy and Traveller communities.[2] 
   
• Between 1 December 2006 and 31 March 2007, ONS undertook a second consultation to establish a more detailed picture of user needs for information on ethnicity, national identity, language and religion.[3]  The consultation received 606 submissions, including from the National Federation of Gypsy Liaison Groups, the Gypsy Council, and the Cardiff Gypsy and Traveller Project. Submissions to this consultation reiterated the need for information on the Gypsy, Irish Traveller and Romany communities, but expressed concern that the inclusion of ‘Romany’ within a Gypsy and Irish Traveller tick-box might “cloud the data-set, making it effectively useless for developing policy with regard to Gypsies and Irish travellers.”
   
• ONS published a paper on the development of a ‘prioritisation tool’ for the selection of ethnicity tick-boxes, in March 2009.[4]   The report concluded that there was strong evidence of a need for data on the Gypsy and Irish Traveller communities.  It also recommended that the term ‘Romany’ be excluded so that ONS could capture the group separately. 

Collecting data

The new Gypsy or Irish Traveller tick box was located under the 'White' heading in the Census questionnaire, as this was where most people from the 'Gypsy or Irish Traveller' group had captured their response in 2001. It was also positioned there to maximise response rates and quality of data.

Figure 1: Census question

Source: ONS

To ensure the highest possible coverage from Gypsy or Irish Travellers, ONS collaborated and engaged with Gypsy and Irish Traveller movements, local government and agencies (further detail on which is provided within Chapter 3 of this note). 

Published data

In January 2014, ONS published an analysis of data collected in the 2011 Census on the characteristics of the Gypsy and Irish Traveller communities in England and Wales.[5] 

The analysis explored the qualifications, economic activity, family relationships, health and accommodation of the communities, and compared their characteristics with other groups.  Key findings included:

• 58,000 people identified themselves as Gypsy or Irish Traveller in the 2011 Census (0.1 per cent of the usual resident population of England and Wales);
   
• People identifying as Gypsy or Irish Traveller had a higher proportion of residents under the age of 20 at 39 per cent. This compares to 24 per cent of the overall England and Wales population who were under 20;
   
• Gypsy or Irish Travellers had the highest proportion with no qualifications1 for any ethnic group (60 per cent) – almost three times higher than for England and Wales as a whole (23 per cent);
• Gypsy or Irish Traveller was the ethnic group with the lowest proportion of respondents who were economically active at 47 per cent, compared to 63 per cent for England and Wales as a whole;
   
• Over half of those who were economically active were employed (51 per cent compared to 75 per cent for the total of England and Wales) and 20 per cent were unemployed (compared to 7 per cent for the whole of England and Wales). Gypsy or Irish Traveller had the highest proportion of self-employed out of the ethnic groups at 26 per cent compared to 14 per cent for England and Wales;
   
• Elementary occupations (such as sales, service or construction) were the most common type of employment at 22 per cent for Gypsy or Irish Traveller (11 per cent for England and Wales as a whole);
   
• Whole house or bungalow was the most common type of accommodation for respondents who identified as Gypsy or Irish Traveller, at 61 per cent (84 per cent for England and Wales as a whole), followed by caravan or other mobile or temporary structure at 24 per cent (0.3 per cent for England and Wales as a whole).

As is noted above, this 2014 analysis explored the characteristics of those who had selected the Gypsy or Irish Traveller tick-box specifically. 

The Committee may be interested to note that, within ‘write-in’ sections of the ethnicity questionnaire, approximately 730 people identified as Roma.[6] 

2.2 Other sources

Surveys

Although the 2011 Census currently offers the most detailed data on GRT communities, ONS also collects some information on these communities within its surveys. 

ONS’s household surveys, including the Crime Survey for England and Wales[7], the Labour Force Survey[8] and the Family Resources Survey[9], all use harmonised (Census) ethnicity questions, which include a tick-box for Gypsy or Irish Traveller groups.  Households are selected to take part in these surveys, at random, from the Postcode Address File.  Council-run hard-standing sites for travellers are listed on the Postcode Address File and are therefore eligible for selection.  Mobile homes are not listed on the Postcode Address File and will not be eligible for selection. 

Very few respondents choose to identify as Gypsy or Irish Traveller within ONS’s household surveys.  Last year only 23 of 88,400 responses to the Labour Force Survey selected the Gypsy or Irish Traveller option.  11 of the 9,000 responses to the Family Resources Survey selected the Gypsy or Irish Traveller option.  No respondents to any other household survey selected the Gypsy or Irish Traveller option. 

Administrative data

As other submissions to the Committee’s inquiry have noted, administrative data (that is, information collected by public bodies in the course of delivering day-to-day services to citizens) do not currently provide a strong source of information on ethnicity.

Education data including School Census data collected by the Department for Education, and Higher Education Statistics Agency data, are the strongest administrative sources available to statisticians on ethnicity.  These sources include ethnic group categories for both ‘Travellers of Irish heritage’ and ‘Gypsy or Roma’.

3. Challenges in collecting data on the Gypsy and Irish Traveller Communities

As other witnesses have highlighted in evidence to the Committee, there are specific challenges associated with collecting data from marginalised communities.  This chapter summarises the challenges that ONS has identified in collecting data on the GRT communities, and what this means for the quality and coverage of official statistics. 

In doing so it focuses in particular on the challenges identified during the conduct of the 2011 Census, when significant work was done to understand the enumeration difficulty of GRT communities. 
 

3.1 Challenges in a Census context

Challenges identified in advance of the 2011 Census

As part of its preparation for the 2011 Census, ONS identified the Gypsy and Irish Traveller communities as groups which were less likely than other groups to return a questionnaire.  Drawing from anecdotal evidence, ONS’s Getting the Count Right document identified the enumeration difficulty of the group as ‘high’.[10] Among the challenges ONS identified, were:

• difficult to locate;
• accommodation variety;
• contact timing;
• mobile population; and
• engagement.

In order to address these challenges, ONS and a range of community groups[11] developed a bespoke set of measures to maximise responses from the Gypsy and Irish Traveller communities (summarised at Annex A).

Gypsies and travellers, like all other population groups, were encouraged to apply for special enumerator roles, as anecdotal evidence had suggested that a familiar face might be more compelling when collecting questionnaires. Legal sites were enumerated by a special enumerator, and local authority gypsy and traveller liaison officers expertise were fed into the collection process. Illegal sites were approached – based on local authority intelligence – prior to census day and questionnaires distributed.  A census leaflet, specific to Gypsies and Travellers, was developed and recruitment postcards were circulated through the Travellers’ Times. 

In addition to the bespoke measures developed for Gypsy and Irish Traveller communities, ONS also prepared easy –read booklets, to maximise responses from groups of lower-literacy levels more generally.

These enumeration procedures were tested during the 2009 Census Rehearsal when a large GRT community in Lancaster was included in the sample.  Good liaison with the local authority and site representatives at this time ensured a good return rate and helped with the design of the 2011 enumeration procedures.

Challenges identified in the course of the 2011 Census

Following the 2021 Census, ONS undertook an evaluation of its engagement with the Gypsy and Traveller communities, which assessed both what went well and what issues arose during Census operations.

Special enumerators reported they were well supported during both delivery and collection phases by early collectors, coordinators and external people/organisations working with GRT communities.  In some cases, local gatekeepers and spokespeople facilitated the smooth running of the enumeration process and even completed the questionnaires on behalf of other residents on the site.  Good cooperation between the central Census publicity team and the special enumeration team had also helped; a press release and feature article had been published in a Roma magazine and a poster, which featured a well-known member of the GRT community, was well received. 

Challenges identified within this evaluation included:

• there was some evidence that poor literacy skills were cited as a reason for not returning responses;
• occupants at at least one site refused to take part because they believed the Census was a ploy to find out the location of Travellers’ sites across the UK in support of a Trespassers Bill; and
• the address list used for the 2011 Census for some is likely to have been inaccurate for some registered and unauthorised sites, because of difficulties ONS faced in establishing a GRT contact at specific sites.  

Limitations of 2011 Census data
 

There are several factors that ONS encourages users to bear in mind when using 2011 Census data on the Gypsy and Irish Traveller communities. 

In the 2011 Census for England and Wales, 58,000 people chose to identify themselves as Gypsy or Irish Traveller.  However, as ONS noted in its 2014 analysis, estimates for the UK from other sources vary between 82,000 to 300,000.[12]

It is important to bear in mind that there are often differences in definitions and coverage of estimates.  For example, some previous estimates for Gypsy or Irish Travellers have included Roma or have been derived from counts of caravans.  The 2011 Census estimate was a measure of those who self-identified as a member of the Gypsy of Irish Traveller communities. 

ONS’s 2014 analysis also highlighted that groups’ concerns about official data collections and/or fear of discrimination may have inhibited a true picture of Gypsy or Irish Travellers in England and Wales being gathered.   Although ONS worked closely with a range of groups to maximise responses to the 2011 Census from the Gypsy and Irish Traveller community, we note (and accept) the evidence provided to the Committee by groups including the London Gypsy and Traveller Unit, suggesting that some members of the Gypsy and Irish Traveller community were reluctant to self-identify in the 2011 Census.

3.2 Challenges in household survey collection

Broadly, the challenges faced by ONS in collecting data on the GRT communities via its household surveys are similar to those it faces in a Census context (outlined above).

There are, however, additional challenges for ONS in collecting robust information about GRT communities via surveys. 

As is noted at section 2.2, because households are selected for survey based on the Postcode Address File, household surveys will not pick-up members of GRT communities living in mobile accommodation. 

Furthermore, while the decennial Census surveys all households, the annual sample size of household surveys (though significant by international standards) does not allow ONS to produce reliable estimates of these communities.  To help address this issue ONS publishes a three-year, combined data-set from the Annual Population Survey.  Drawing from this combined dataset it has been possible to publish limited analysis of, for example, the personal wellbeing of Gypsy and Irish Traveller communities.[13]     

4. Developing the evidence base

ONS’s Strategy, Better Statistics, Better Decisions, sets out our commitment to respond to unmet statistical needs, using innovative, efficient and effective methods.  In the context of developing official statistics on GRT communities, ONS has already begun preparations for the 2021 Census and is working to support the Government’s Race Disparity Audit.  This chapter briefly sets out the work underway in these fields relating to data on the GRT communities. 

4.1 2021 Census

ONS has already begun preparations for the next Census, due to be held in 2021.  In June 2015 ONS launched a public consultation asking users for their views on the topics that the 2021 Census questionnaire in England and Wales might cover.

ONS published its response to the consultation in May 2016.[14]  In its response, ONS noted that it proposed to continue to collect information about ethnic group and national identity in the 2021 Census and an ‘Ethnic group stakeholder follow-up survey’ is now being conducted, to develop a more detailed picture of user requirements. 

Ahead of the 2021 Census, ONS will begin developing a bespoke set of measures to maximise responses to the 2021 Census from GRT communities.  In doing so, ONS will be drawing from the lessons it identified during the conduct of the 2011 Census, and working closely with community groups.  We would welcome the opportunity to keep members of the Committee informed of our preparations ahead of the 2021 Census. 
 

4.2 Race Disparity Audit

On 27 August 2016, the Prime Minister announced an audit to tackle racial disparities in public service outcomes. The audit will:

1. review Government data to identify racial inequalities in outcomes from contact with public services, and any gaps in data collection; and
2. publish all data in a single place so that the public can search the data to show inequalities in outcomes by geography, age or social-economic category.

The aim of the audit is to ensure a clearer focus on overall outcomes and experiences, exposing the multiple disadvantages faced by ethnic minority groups across society, with transparency driving change.  

As part of this programme of work, ONS is currently working closely with colleagues across Government to identify sources and data gaps, including on GRT communities.

March 2017 

Annex A