Social model of disability: discovery phase interim report (summary)

This paper reports findings from the first phase of a research project to develop survey questions, response options, and guidance that reflect the social model of disability. This stage (the discovery phase) comprised of 3 data collection exercises: a systematic review of existing survey questions, workshops with users of disability data, and deliberative workshops with disabled people.

The Welsh Government adopted the social model of disability in 2002 making Wales one of the first countries in the world to do so. However, data collection is still heavily influenced by the Equality Act (2010) (UK legislation) and, by extension, the medical model of disability. Welsh Government have identified that new questions are required to ensure the social model is reflected in its actions, data collection, and decision-making, and in the delivery of all services across Wales.

The National Centre for Social Research (NatCen) were therefore commissioned to develop and test a new set of demographic survey questions, response options, and guidance to align with the social model of disability. The questions and guidance will be designed to: support a diverse range of accessibility needs; work across different contexts and languages, including Welsh and British Sign Language (BSL); and be suitable across various data collection modes, such as online and in-person formats.

The research is split into 2 phases: an initial discovery phase and a subsequent ‘alpha’ phase, which will develop and test new survey questions. This report synthesises findings from the first (discovery) phase, the aims of which were to:

• understand the different models of disability and how these models are reflected in the survey questions used to generate government statistics
• understand the different ways data users currently use survey data to generate statistics related to disabled people, or people with impairments, and what is currently lacking in the existing approach
• identify the range of survey questions, response options, and guidance currently being used to collect data on disabled people while reflecting the social model of disability
• understand disabled peoples’ perspectives on how information about impairments should be collected and what type of questions should be asked in surveys

The 3 data collection methods used in the discovery phase

Workshops with data users

Three workshops were conducted with people that use data and statistics about disabled people. Each workshop was scheduled for 2 hours and included between 8 to 16 participants. These included government researchers, policy officials, and statisticians; public sector bodies employees; and representatives from third sector and community organisations who act as advocates for disabled people.

Review of existing measures

A comprehensive systematic review was undertaken of existing survey measures that captured information regarding disabled people and their impairments. In total, question sets from 20 surveys, from the UK and other countries, were reviewed. The review focused on identifying and evaluating the extent to which these measures met, or could be adapted to meet, the needs of data users and whether they align with language of the social model of disability.

Deliberative workshops with disabled people

Deliberative research methods were used to conduct 2 workshops with disabled people from across Wales. The workshops were split into 2 half-day sessions and consisted of presentations, whole group discussions, and smaller break-out discussions. Across the 2 sessions, 40 people participated in the workshops, representing a breadth of perspectives and lived experience.

Data user workshops

When asked about current data sources, participants most commonly reported using internal organisational surveys and government surveys at local, national, and international levels to collect data on impairment and disabled people. All participants reported using these data to monitor and assess trends or evaluate and improve circumstances for disabled people. Some data users also reported using this information to evaluate workforce diversity, while others (particularly data users from the third sector and community organisations) used data to further understand the needs of, and advocate for, communities they represent.

Data users were also asked to reflect on the benefits and limitations of the disability harmonised standards produced by the Government Statistical Service (GSS). The main advantage identified by participants was the standardisation of questions, allowing for the comparability of data over time and across different surveys. In terms of limitations, participants felt the harmonised questions were based on a “deficit model” and included “negative” language. They could also create barriers to engagement or exclude certain groups, for example, people with fluctuating conditions, neurodivergent people, and first-language British Sign Language users.

Data users engaged positively with the prospect of using social model-informed measures. They discussed needing a shift of focus away from viewing ‘disability’ as a problem with the individual, towards a focus on societal and environmental barriers, and that a “linguistic shift” was needed. Using free-text response options would allow respondents to define their own impairment, capture more in-depth data, and give more scope for respondent self-identification.

However, data users also stated that new questions based on the social model would require guidance and may increase the length of surveys, creating a more taxing user experience. They emphasised the need to strike a balance between survey usability and how much data could be captured. It was noted that the format and mode of questionnaires may still be exclusionary, which must be addressed by using a range of modes and formats to ask and answer the questions. These could include interviewer assisted in-person surveys and self-complete online surveys, and having easy read and spoken formats, and surveys in British Sign Language, for example.

Across all 3 workshops, participants felt the questions should be redesigned with input from the people they are intended to serve, and not just organisations and professionals. Suggestions were made for retaining the existing questions and supplementing them with new measures.

Systematic review of existing survey measures

Overall, the review of 20 existing measures suggests that, whilst some partially align with the social model of disability, most would require substantial adaptation to fully reflect its principles and meet the needs of data users. The biopsychosocial model of disability was the conceptual framework used for most existing surveys.

There was limited consideration across most surveys of how fluctuating or progressive conditions affect people's daily lives, and most of the reviewed surveys were limited in their ability to capture data on environmental and social barriers. Even in cases where data relating to barriers were collected, they tended to be treated as supplementary topics rather than integrating them fully into the conceptualisation of disability.

Many surveys avoided using the word ‘disability’ directly and instead preferred to ask about ‘long-term health conditions’ or ‘activity limitations’. Some surveys used language that is now considered outdated or insensitive, and others used language which conflated impairments with medical symptoms, further reducing the focus away from the social model of disability.

Many surveys were originally designed for interviewer-administered face to face formats and have only more recently been adapted for self-completion, telephone, or online administration. Also, while many surveys permitted proxy responses for children or individuals unable to complete the survey independently, surveys providing alternative accessible formats were rare. This presents a potential barrier to inclusivity, particularly for disabled people who may face communication, sensory, or cognitive impairments and/or barriers.

Deliberative research with disabled people

During the deliberative workshops, participant discussions included conceptual and more practical reflections on ideas like inclusivity and comparability, language use, and assessments of different survey questions (both pre-existing and potential new questions).

Language featured in discussions in relation to how participants identified themselves, and how they perceived and understood various survey questions. There was no consensus on preferred terminology, but many described a distinction between being disabled and having impairments, a distinction which is also central to the social model. Most participants held a negative view of the terms, ‘disability’ and ‘being disabled’, associating these with restrictions and negative societal perceptions. Some also suggested that the word ‘barriers’ was too abstract to be fully understood by the general public and may be associated exclusively with physical barriers.

Participants had mixed views on the importance of comparability. Some stated it was an important consideration in developing survey questions, but others found the concept of data comparability confusing and often conflated it with the use of simple, clear, and accessible language in survey questions.

Most participants judged inclusivity to be more important than comparability when designing survey questions. However, inclusivity was referred to in different ways: either to ensure different impairments or experiences of being disabled were captured in survey questions, or as shorthand for inclusive language. Some participants said offensive or dated wording (for instance, the use of the term ‘disorder’) would make people reluctant to respond to questions. A minority of participants felt that the idea of inclusive language was becoming too complex, and people were too easily offended.

When reflecting on the ‘core’ harmonised standards that measure long-lasting health conditions and illness (GSS), and activity restrictions (GSS), initially, many participants said they would find these questions easy to answer. However, when these participants reflected on the questions and discussed them with others, many began to identify numerous drawbacks, concluding that they are difficult to answer and exclusionary. Some participants said they were exclusionary as they did not capture the experiences of some disabled people, had exclusionary and negative language, and the responses of ‘a little’ or ‘a lot’ were subjective.

In comparison to the ‘core’ harmonised questions, the impairment harmonised standard (GSS) was viewed more positively as it was seen to display a better understanding of impairments and felt “gentler” than the ‘core’ standards. Participants suggested that the inclusion of an ‘Other’ response and the option to select ‘all the answers that apply’ was beneficial as it enabled the collection of more in-depth and accurate data that reflects disabled people’s complex experiences. However, when considering the question in isolation, participants overwhelmingly felt that the impairment question is an improvement on the ‘core’ standards but needs further improvement.

Most participants reacted positively to the idea of a social model-inspired survey question. Some participants said the sample question put to them by facilitators was “empowering” and would make them feel like the government “cared” for them. However, many participants also said it was too lengthy, complicated, and wordy. Most participants were in favour of the harmonised impairment question and a social model-style question being asked together, as this could help link data on impairments with data on barriers and produce more actionable data.

No single existing question-set, that met the needs of all groups, was identified during the discovery phase. Participant views on future question development were also mixed, with some preferring to build upon the harmonised question wording for time-series analysis (and for alignment with Equality Act 2010 definitions), whereas others wanted to prioritise the development of entirely new questions, embedding a social model approach with more appropriate and inclusive language.

Based on the findings from the discovery phase, the following options for questionnaire development are recommended to be prioritised during the next project phase. These options will not be mutually exclusive.

1. An updated impairment question, based on the GSS harmonised impairment standard.
2. A new question on societal barriers that people with impairments encounter.

For option 1, it is recommended that the design work should focus on fine-tuning the existing harmonised impairment question to address some of the barriers and limitations identified in the discovery phase. These included the use of outdated and exclusionary terminology; phrasing of conditions that affect people ‘socially’ or ‘behaviourally’; clarity over fluctuating conditions; and suitability for use by parents who answer about their children.

For option 2, it is recommended that a single follow-up question is produced for those reporting one or more impairments in response to the previous question. The follow-up question should focus on barriers, asking respondents to identify any barriers they have experienced in relation to their impairments from a ‘check-all that apply’ list of possible options.

Question development work for option 2 should focus on developing a high-level list of barriers that will be relevant to both policy makers and to a broad range of disabled people. It is noted that a single question will limit the granularity of information on all types of barriers encountered.

For both options, questions should be designed and tested for use in different completion modes. This will allow for the accommodation of different formats as surveys continue to transition from interviewer administration toward web-first or self-completion designs.

The priority of the questionnaire development would be to ensure they meet the needs of data users and to work with disabled people to develop a clear list of the barriers they feel are most pressing and most relevant to different types of impairment.

Report author: Lisa Rutherford, Harry Pearse, Nathan Hudson, Jo d’Ardenne, Aditi Das, Alessandra Radicati, Yasmin Begum

Views expressed in this report are those of the researchers and not necessarily those of the Welsh Government.

For further information please contact:
Laura Tolland
Disability Disparity Evidence Unit
Equality, Poverty, Children’s Evidence and Support Division
Welsh Government
Cathays Park
Cardiff
CF10 3NQ

Email: EqualityEvidenceUnit@gov.wales

Social research number: 78/2025
Digital ISBN: 978-1-80633-148-2