Judith BaartpersName, Willem ElberspersName, Alice SchipperspersName, Sarah PolackpersName (2025date). “The Definition Dilemma: How Definitions of Disability Shape Statistics on Social Participation..” International Journal of Environmental Research and Public Healthtitle·j.
Object of study — how the choice of disabilityterm·key measure shapes statistics on the participationterm·key of persons with disabilities
Research question — to what extent the measured participationterm·key of persons with disabilities varies with the disabilityterm·key measure used
Contribution — evidence that different disabilityterm·key measures identify different people and produce different participationterm·key gaps
phenomenon · claim-type · cross-domain-transfer · mechanism · claim-relation · replication-status · limitation — grounded in the full text| phenomenon | the absence of a universally agreed definition of disability — “Yet there is no universally agreed definition for disability: its meaning varies and is dependent on the context in which it is created [ 8 , 9 , 10 , 11 , 12 ].” | |
| claim-type | existence — “The findings demonstrate that there is no single, uniform category of “persons with disabilities” and that this has implications for subsequent outcomes on indicators of participation.” | |
| method — evidence | — “For this study, we carried out secondary analysis using an existing data set from a population-based survey on disability amongst Syrian refugees in Sultanbeyli district in Istanbul in 2019, carried out by author S.P. and colleagues.” | |
| object of study — evidence | — “This paper explores how the choice of disability measure influences findings on inclusion indicators of persons with disabilities, in this case in a refugee setting.” | |
| research question — evidence | — “This study aims to understand to what extent the level of participation of persons with disabilities varies depending on the measure of disability used in research.” | |
| contribution — evidence | — “This study provides evidence that participation measures differ depending on the measure of disability used.” | |
| sense | Inclusion (common usage) - the opposite of social exclusion — “Or, more commonly, “the opposite of social exclusion” (ibid).” | senses[0] |
| sense | disability (this paper) - complex and multi-dimensional, with a definition that may change with its purpose — “Disability is complex, challenging and multi-dimensional, making it difficult to define, and the definition may need to change depending on its purpose [ 13 ].” | senses[1] |
| relation | UN Convention on the Rights of Persons with Disabilities —same-as→ UNCRPD — The UN Convention on the Rights of Persons with Disabilities (UNCRPD) [ 2 ] and the Sustainable Development Goals [ 3 ], emphasise that persons with disabilities should have the same access to services and opportunities in society as those without. — “The UN Convention on the Rights of Persons with Disabilities (UNCRPD) [ 2 ] and the Sustainable Development Goals [ 3 ], emphasise that persons with disabilities should have the same access to services and opportunities in society as those without.” | concept_edges[0] |
| weight | UN Convention on the Rights of Persons with Disabilities —same-as→ UNCRPD — 0.9 — “The UN Convention on the Rights of Persons with Disabilities (UNCRPD) [ 2 ] and the Sustainable Development Goals [ 3 ], emphasise that persons with disabilities should have the same access to services and opportunities in society as those without.” | concept_edges[0] |
| relation | Washington Group Short Set —same-as→ WGSS — For the purposes of the current analysis, we compared three measures of disability collected in the survey: the Washington Group Short Set (WGSS), the Washington Group Short Set—Enhanced (WGE) and a two-question sequence asking about limitations due to disability or illness, herein referred to as the “two-question set” (2QS) for brevity. — “For the purposes of the current analysis, we compared three measures of disability collected in the survey: the Washington Group Short Set (WGSS), the Washington Group Short Set—Enhanced (WGE) and a two-question sequence asking about limitations due to disability or illness, herein referred to as the “two-question set” (2QS) for brevity.” | concept_edges[1] |
| weight | Washington Group Short Set —same-as→ WGSS — 0.9 — “For the purposes of the current analysis, we compared three measures of disability collected in the survey: the Washington Group Short Set (WGSS), the Washington Group Short Set—Enhanced (WGE) and a two-question sequence asking about limitations due to disability or illness, herein referred to as the “two-question set” (2QS) for brevity.” | concept_edges[1] |
| relation | two-question set —same-as→ 2QS — For the purposes of the current analysis, we compared three measures of disability collected in the survey: the Washington Group Short Set (WGSS), the Washington Group Short Set—Enhanced (WGE) and a two-question sequence asking about limitations due to disability or illness, herein referred to as the “two-question set” (2QS) for brevity. — “For the purposes of the current analysis, we compared three measures of disability collected in the survey: the Washington Group Short Set (WGSS), the Washington Group Short Set—Enhanced (WGE) and a two-question sequence asking about limitations due to disability or illness, herein referred to as the “two-question set” (2QS) for brevity.” | concept_edges[2] |
| weight | two-question set —same-as→ 2QS — 0.9 — “For the purposes of the current analysis, we compared three measures of disability collected in the survey: the Washington Group Short Set (WGSS), the Washington Group Short Set—Enhanced (WGE) and a two-question sequence asking about limitations due to disability or illness, herein referred to as the “two-question set” (2QS) for brevity.” | concept_edges[2] |
| replication | original | |
| limitation | Participation was measured only by schooling, marital status and employment, a narrow subset of inclusion. — “The measures of participation were restricted to schooling, marital status, and employment, which, while important, capture only a narrow subset of societal inclusion.” | limitations[0] |
| limitation | Results could not be broken down by age, sex and type of impairment because the sample was small. — “Data disaggregation by age, sex and type of impairment was not possible in our study due to limited sample size.” | limitations[1] |
| limitation | Mental health conditions in the refugee population may have shaped how physical and sensory functioning was reported. — “The prevalence of mental health conditions, including anxiety and depression, within the refugee population studied may have influenced how individuals perceived and reported physical and sensory functioning.” | limitations[2] |
finding · quantity · variable · definition · key-term — grounded in the full text| finding | Disability prevalence ranged from 13.5% to 22.4% depending on the measure used. — “The prevalence of disability ranged from 13.5 to 22.4% depending on the measure used.” | content_tags.findings[0] |
| finding | Only 8.8% of respondents counted as disabled under both the WGSS and the 2QS. — “Only 8.8% are categorised disabled by both measures, i.e., for disability statistics they would be considered a person with a disability regardless of the disability measurement used.” | content_tags.findings[1] |
| quantity | 1555 — — “In total, 1555 respondents aged 18+ years were included in the analysis.” | content_tags.quantities[0] |
| key-terms | disability, participation, inclusion, Washington Group questions, UNCRPD, Sustainable Development Goals, refugees, disability statistics, prevalence, social exclusion |
| concept | Inclusion — a desired goal that requires equality of opportunity and participation in the rudimentary and fundamental functions of society — “Inclusion is best defined as “a desired goal that requires equality of opportunity and participation in the rudimentary and fundamental functions of society” [ 4 ] (p. 35).” | concepts[0] |
| concept | UN Convention on the Rights of Persons with Disabilities — not defined in this paper's own words; the United Nations treaty on the rights of disabled people — “The UN Convention on the Rights of Persons with Disabilities (UNCRPD) [ 2 ] and the Sustainable Development Goals [ 3 ], emphasise that persons with disabilities should have the same access to services and opportunities in society as those without.” | concepts[1] |
| concept | UNCRPD — UN Convention on the Rights of Persons with Disabilities — “The UN Convention on the Rights of Persons with Disabilities (UNCRPD) [ 2 ] and the Sustainable Development Goals [ 3 ], emphasise that persons with disabilities should have the same access to services and opportunities in society as those without.” | concepts[2] |
| concept | International Classification of Functioning, Disability and Health — the interaction between the underlying health condition/impairment and contextual factors — “The International Classification of Functioning, Disability and Health (the interaction between the underlying health condition/impairment and contextual factors) [ 14 , 15 ] developed by the WHO has, since its development, provided a common framework for conceptualising disability, but needs operationalising.” | concepts[3] |
| concept | Washington Group Short Set — not defined in this paper's own words; a short survey module asking how hard basic activities are for a person — “For the purposes of the current analysis, we compared three measures of disability collected in the survey: the Washington Group Short Set (WGSS), the Washington Group Short Set—Enhanced (WGE) and a two-question sequence asking about limitations due to disability or illness, herein referred to as the “two-question set” (2QS) for brevity.” | concepts[4] |
| concept | WGSS — Washington Group Short Set — “For the purposes of the current analysis, we compared three measures of disability collected in the survey: the Washington Group Short Set (WGSS), the Washington Group Short Set—Enhanced (WGE) and a two-question sequence asking about limitations due to disability or illness, herein referred to as the “two-question set” (2QS) for brevity.” | concepts[5] |
| concept | two-question set — a two-question sequence asking about limitations due to disability or illness — “For the purposes of the current analysis, we compared three measures of disability collected in the survey: the Washington Group Short Set (WGSS), the Washington Group Short Set—Enhanced (WGE) and a two-question sequence asking about limitations due to disability or illness, herein referred to as the “two-question set” (2QS) for brevity.” | concepts[6] |
| concept | 2QS — two-question set — “For the purposes of the current analysis, we compared three measures of disability collected in the survey: the Washington Group Short Set (WGSS), the Washington Group Short Set—Enhanced (WGE) and a two-question sequence asking about limitations due to disability or illness, herein referred to as the “two-question set” (2QS) for brevity.” | concepts[7] |
Full text — tagged
The Definition Dilemma: How Definitions of Disabilityterm·key Shape Statistics on Social Participationterm·key.
Abstract To monitor progress in including persons with disabilities, including in interventions which can improve their health and quality of life, it is crucial to collect data on their participationterm·key. However, there are many different ways of defining disabilityterm·key and thus categorising individuals as disabled/not disabled, which may impact measures of participationterm·key. This study aims to assess the relationship between three different measures of disabilityterm·key and the level of participationterm·key of persons with disabilities. We analysed data from a population-based surveyterm·mth of disabilityterm·key amongst Syrian refugeesterm·key in Istanbul, which collected data on disabilityterm·key using the Washington Group enhanced set of disabilityterm·key questions as well a two-question setterm asking about perceived activity limitations due to disabilityterm·key. The prevalenceterm·key of disabilityterm·key ranged from 13.5 to 22.4% depending on the measure used. The group of people that are categorised as disabled also differs, indicating who is being seen as disabled changes when a different measure is used. Levels of participationterm·key, with regards to paid work, education and being partnered, also varied by measure, for example, being in paid work ranged from 26% to 38%. These findings underscore the importance of carefully selecting and clearly defining disabilityterm·key measures in studies, (health) interventions and policy contexts.
Keywords: disabilityterm·key, Washington group questionsterm·key, participationterm·key, UNCRPDterm·dom, Sustainable Development Goalsterm·key, inclusionterm·dom, analysis 1. Introduction “We say that 16% of the world’s population has a disabilityterm·key. The majority of them do not use the word “disabilityterm·key” to describe themselves. To understand disabilityterm·key requires us to make space for diverse perspectives and diverse ways people describe their own experiences” [ 1 ].
The UN Convention on the Rights of Persons with Disabilitiesterm·dom (UNCRPDterm·dom) [ 2 ] and the Sustainable Development Goalsterm·key [ 3 ], emphasise that persons with disabilities should have the same access to services and opportunities in society as those without. This means that the international development sector is expected to promote the inclusionterm·dom of persons with disabilities in their policies and programmes, including in interventions aimed at benefiting health and quality of life. To understand the extent to which this is being achieved requires monitoring the participationterm·key of persons with disabilities. This requires understanding and defining who persons with disabilities are. However, different approaches are used to measure and define disabilityterm·key, which in turn are likely to influence the levels of participationterm·key that are measured and indicators of progress towards inclusionterm·dom. This study aims to understand to what extent the level of participationterm·key of persons with disabilities varies depending on the measure of disabilityterm·key used in research.
The UNCRPDterm·dom notes as its general principle “full and effective participationterm·key and inclusionterm·dom in society” [ 2 ] (p. 5). The Sustainable Development Goalsterm·key (SDGs) also highlight the importance of the inclusionterm·dom of persons with disabilities, noting its goal to “empower and promote the social, economic and political inclusionterm·dom of all, irrespective of age, sex, disabilityterm·key, race, ethnicity, origin, religion or economic or other status” [ 3 ] (p. 21). Inclusionterm·dom is best defined as “a desired goal that requires equality of opportunity and participationterm·key in the rudimentary and fundamental functions of society” [ 4 ] (p. 35). Or, more commonly, “the opposite of social exclusionterm·key” (ibid). Policymakers need to collect data on persons with disabilities and monitor their participationterm·key to know whether they are achieving inclusionterm·dom in their policies and programmes.
Societal participationterm·key of people with disabilities (i.e., are those with disabilities partaking and being served on an equal basis) is a key measure to track progress regarding the inclusionterm·dom of persons with disabilities. As noted by the Washington Group on Disability Statisticsterm·key, “If there are statistical differences in the number of people with disabilityterm·key in school or work, marrying or voting, then we can begin to discuss ‘exclusion’” [ 5 ] (p. 3). Participationterm·key of persons with disabilities and progress on the SDG targets can thus be assessed by disaggregating outcome indicators—such as employment, schooling, or access to health care. “These participationterm·key levels should be equal; those with disabilities should be participating equally in society compared to those without disabilities. If the levels are not equal, appropriate accommodations have not been made” [ 6 ] (p. 1169). This is similar to Rimmerman’s explanation of how social inclusionterm·dom or exclusion is regularly measured, by “comparing the status of the excluded subpopulation with the general public by using multiple indicators” [ 4 ] (p. 42); comparing groups of people with and without disabilities can help us to “better understand the equalization of opportunity” [ 7 ] (p. 2).
Yet there is no universally agreed definition for disabilityterm·key: its meaning varies and is dependent on the context in which it is created [ 8 , 9 , 10 , 11 , 12 ]. Disabilityterm·key is complex, challenging and multi-dimensional, making it difficult to define, and the definition may need to change depending on its purpose [ 13 ]. The International Classification of Functioning, Disability and Healthterm·dom (the interaction between the underlying health condition/impairment and contextual factors) [ 14 , 15 ] developed by the WHO has, since its development, provided a common framework for conceptualising disabilityterm·key, but needs operationalising. Commonly used disabilityterm·key definitions result from different disabilityterm·key models, such as the functional (the inability to do something); the administrative (the individual has been processed as disabled by the welfare state); the subjective (the individual perceives him/herself as disabled); the social (barriers in society keep people with impairments from participating in society) [ 16 ] and the medical (disabilityterm·key is a deviation from the normal functioning of the body). More models and ensuing definitions exist globally, as ways of constructing disabilityterm·key come from the cultural and social context within which they exist [ 11 ].
The definition and subsequent assessment tools used to measure disabilityterm·key shapes who are labelled, categorised and recognised as having a disabilityterm·key. Our previous analysis comparing functional and subjective assessment tools of disabilityterm·key found that there was a core group of people who would be categorised as disabled using either assessment tool, but also people whose categorisation would depend on the tool used [ 17 ] (p. 10). Other studies have found poor agreement between functional assessments of disabilityterm·key (such as the Washington Group questionsterm·key) with direct self-identification questions (such as “do you have a disabilityterm·key?”), with some studies finding less than 50% of respondents being categorised as “disabled” using both methods [ 7 , 17 , 18 , 19 , 20 , 21 , 22 ]. A comparison of a direct self-identification question of disabilityterm·key with administrative data (whereby students that had an Individual Education Plan were classified as disabled) found 89% of respondents being labelled as “disabled” using both methods, concluding that individuals considered the two concepts to be distinct constructs [ 23 ]. Studies comparing a clinical assessment of impairment to a self-reported functional measure also demonstrate differences in populations categorised as disabled [ 24 , 25 , 26 ]. One study, comparing four different disabilityterm·key measures found prevalenceterm·key estimates ranging from 5.6% to 29.7% [ 27 ]. In summary, these studies indicate that “The different approaches taken […] result in the identification of substantially different, though overlapping, groups as disabled, and a range of different estimates of disabilityterm·key prevalenceterm·key” [ 21 ] (p. 23).
Given that each method of measuring disabilityterm·key categorises different individuals as disabled, the process of choosing which method of categorisation is used is a political process that can potentially have far-reaching effects on the lives of individuals, such as whether or not they are eligible for social protection programmes. As noted by Beaudry, “once a definition of disabilityterm·key is used in legal contexts and social movements or embedded in a policy, it can do both harm and good. For instance, it can contribute to marginalising people with disabilities or it can empower them to make claims through their right to equality” [ 28 ] (p. 4). It is thus crucial to understand that labelling is a social construct dependent on who defines which labels are used and why; it is never apolitical [ 29 ].
We thus know from earlier research that the prevalenceterm·key of disabilityterm·key is dependent on the measure of disabilityterm·key used: different tools identify different people as disabled. What has scarcely been studied yet and has not been studied in the international development context, is how differing disabilityterm·key measures would influence levels of participationterm·key, given that inclusionterm·dom is assessed by comparing key outcomes of those with disabilities to those without disabilities. If participationterm·key measures change with changing disabilityterm·key assessment tools, then without thoughtfully addressing how disabilityterm·key is measured, and how individuals are categorised as disabled, information on the progress achieved in promoting inclusionterm·dom may be skewed—as well as any assessments on the effectiveness of policies and programmes that target and promote inclusionterm·dom.
This paper explores how the choice of disabilityterm·key measure influences findings on inclusionterm·dom indicators of persons with disabilities, in this case in a refugeeterm·key setting. Refugeeterm·key populations offer a particularly relevant case for examining how different definitions of disabilityterm·key shape who is identified as disabled. Given their diverse experiences of displacement, trauma, and disrupted access to care, refugeesterm·key are likely to present a wide range of functional limitations and psychological challenges. In such heterogeneous settings, the boundaries of disabilityterm·key are especially fluid and may shift significantly depending on the measurement tool used. This makes them a compelling population for exploring how definitional choices influence both categorisation and the way we assess inclusionterm·dom or exclusion.
We compare three measures of disabilityterm·key: the Washington Group Short Setterm (WGSSterm), the Washington Group Short Setterm—Enhanced (WGE), which ask about the level of difficulty in different functional domains (e.g., seeing, hearing, self-care) and a two-question sequence asking about limitations due to disabilityterm·key or illness. The Washington Group questionsterm·key and questions about limitations due to disabilityterm·key or illness are commonly used in censuses and surveysterm·mth to generate disability statisticsterm·key, but few studies have compared them, particularly within humanitarian contexts and conflict-affected populations. We assess the level of agreement and compare key indicators of inclusionterm·dom (education, labour market participationterm·key and marital status) between people with and without disabilities categorised according to each of these disabilityterm·key measures.
2. Materials and Methods For this study, we carried out secondary analysis using an existing data set from a population-based surveyterm·mth on disabilityterm·key amongst Syrian refugeesterm·key in Sultanbeyli district in Istanbul in 2019, carried out by author S.P. and colleagues. Detailed methods are presented in a previous publication [ 30 ]. Briefly, multi-stage cluster sampling was used to select people for participationterm·key in the surveyterm·mth. Participants were visited in their households for an interview. In total, 4018 people aged 2+ years were enumerated, of which 3084 took part in the surveyterm·mth. Analysis for this paper was restricted to surveyterm·mth respondents aged 18 and older, leaving us with 1555 respondents to include in this study. This was because the Washington Group Questionsterm·key used for children are different.
Data were collected on Android tablets with Open Data Kit software (Available online: https://getodk.org/ accessed on 7 April 2025). Ethical approval was obtained from Istanbul Sehir University, Republic of Turkey Ministry of Interior: Directorate General of Migration Management and the London School of Hygiene and Tropical Medicine. Informed consent was obtained from all participants. Those that were identified as having health needs were referred for further services. Data collected included socio-demographic and economic data including age, sex, education, marital status and employment.
For the purposes of the current analysis, we compared three measures of disabilityterm·key collected in the surveyterm·mth: the Washington Group Short Setterm (WGSSterm), the Washington Group Short Setterm—Enhanced (WGE) and a two-question sequence asking about limitations due to disabilityterm·key or illness, herein referred to as the “two-question setterm” (2QSterm) for brevity. Table 1 presents more details of each of these measures and the criteria used to classify people as disabled for the purposes of disability statisticsterm·key.
Disabilityterm·key measures used in the research.
We compared key indicators between persons with and persons without disabilities and assessed whether this differs depending on the measure of disabilityterm·key used. The measures of participationterm·key collected in this study were level of education, paid labour market participationterm·key and marital status. Analysis was completed using IBM SPSS Statistics 29. We calculated the prevalenceterm·key of disabilityterm·key (with 95% confidence intervals) according to each measure, stratified by age and sex. The degree of overlap between each of the two WGQ sets with the 2QSterm was visualised using Venn diagrams. Odds ratios were calculated to compare participationterm·key indicators between persons with and without disabilities, with three separate analyses, one for each disabilityterm·key measure.
3. Results 3.1. Prevalenceterm·key of Disabilityterm·key Measures In total, 1555 respondents aged 18+ years were included in the analysis. Just over a third of the sample (40.5%) were aged 18–29 years old; only 5.1% were 60+ years old. Slightly more females (56.7%) than males (43.3%) were included in the sample.
Overall, 13.5% (95% CI 11.8–15.3; n = 210) were categorised as having a disabilityterm·key using the WGSSterm, 21.7% (95% CI 19.6–23.8; n = 337) using the WG-E and 22.4% (95%CI 20.4–24.6; n = 349) based on the 2QSterm. The number of people in the population that would be considered disabled thus increases considerably when the 2QSterm or WGE measure is used. The prevalenceterm·key of disabilityterm·key increases with age, regardless of the measure used. Prevalenceterm·key was higher for women using the Washington Group measures, and higher for men using the 2QSterm, but these differences were not statistically significant ( Table 2 ).
Prevalenceterm·key of disabilityterm·key in the population: WG short set, WG Enhanced and 2QSterm, by sex and age group.
3.2. Overlap and Differences Between Disabilityterm·key Measures In Figure 1 and Figure 2 we present the overlap of the individuals categorised as disabled using the Washington Group measures and the individuals labelled as disabled using the 2QSterm measure. Figure 1 compares the WGSSterm with 2QSterm.
Percentage of respondents categorised as disabled by 2QSterm and WGSSterm.
Percentage of respondents categorised as disabled by 2QSterm and WGE.
Looking at who makes up these groups: 4.7% of the full study population is categorised disabled using the WGSSterm but would be considered not disabled according to the 2QSterm measure, 13.6% are categorised disabled using the 2QSterm measure, but would be considered as not disabled according to the WGSSterm. Only 8.8% are categorised disabled by both measures, i.e., for disability statisticsterm·key they would be considered a person with a disabilityterm·key regardless of the disabilityterm·key measurement used.
Comparing the WGE with the 2QSterm measure ( Figure 2 ), 10.1% of the respondents are categorised as having a disabilityterm·key using the WGE but not the 2QSterm, 10.9% are categorised as having a disabilityterm·key using the 2QSterm but do not have a disabilityterm·key according to the WGE, 11.6% is categorised as having a disabilityterm·key using both the WGE and the 2QSterm measure.
Of all the respondents categorised as disabled using the WGSSterm or the WGE, older people are more likely to also be categorised as disabled using the 2QSterm (81.3% of 60+ year-olds would be considered disabled by both WGSSterm and 2QSterm; whereas only 39% of the 18–29-year-olds that are disabled using WGSSterm are also disabled using 2QSterm) ( Table 3 ). That is, of the people who report difficulties in functioning, older people are more likely to also identify themselves as having a disabilityterm·key or illness that limits their activities.
Demographic characteristics of persons categorised as disabled by both WGSSterm or WGE and 2QSterm.
3.3. Measures of Participationterm·key Using Different Definitions of Disabilityterm·key Table 4 presents the three measures of participationterm·key, disaggregated by the three different measures of disabilityterm·key used in this study.
Percentage and odds ratios of not having had any schooling, being married/living together and doing paid work, by different measures of disabilityterm·key.
The percentage of respondents categorised as disabled who have not attended any schooling varies from 13% (2QSterm) to 14% (WGE) to 18% (WGSSterm). Regardless of the measure, the odds of having no formal education were significantly higher for people with disabilities; but the magnitude of the effect size varied slightly (WGSSterm: OR 2.53, 95% CI 1.69–3.78); 2QSterm: OR 1.82, 95% CI 1.26–2.63; WGE: OR 1.66, 95% CI 1.15–2.41).
Persons with disabilities were consistently less likely to be currently partnered (i.e., married or living together) than persons without disabilities, although this difference was only statistically significant using the WGE (OR 0.77 (95% CI 0.54–0.97).
For paid work (defined as having regular or irregular paid work or seeking paid work) we restricted the analysis to participants who identified as male, as very few women (<5%) in the study population were in paid work. Approximately a quarter (26%) of men with disabilities participate in paid work using the WGSSterm measure of disabilityterm·key, as compared to 35% using the 2QSterm and 38% using the WGE. Regardless of the measure, men with disabilities were significantly less likely than men without disabilities to be in paid work, but the magnitude of the difference was greatest when using the WGSSterm (OR 0.28, 95% CI 0.18–0.44).
4. Discussion In order to understand whether the inclusionterm·dom of persons with disabilities is being achieved, and to operationalise the ICF approach towards disabilityterm·key, requires the use of tools to measure disabilityterm·key in censuses and surveysterm·mth. Using three different measures of disabilityterm·key on one population—in this case Syrian refugeesterm·key in Turkey—enables us to analyse how research results change with varying definitions and ensuing measures of disabilityterm·key. First, we note that the disabilityterm·key prevalenceterm·key in the population changes with each measure. This corroborates earlier research and general disability statisticsterm·key [ 7 , 17 , 18 , 19 , 20 , 21 , 22 , 23 ]. Thus, any research, census or statistics that publish data on persons with disabilities should always note which measure and model of disabilityterm·key was used.
In addition, we see that the characteristics of those categorised as disabled change depending on the measure used: older people who are considered disabled using a functional measure of disabilityterm·key are more likely than younger people to also be considered disabled using a limitations measure of disabilityterm·key; those with mental health conditions such as anxiety or depression are more likely to be categorised as disabled using the WGES and the 2QSterm than the WGSSterm. People can thus experience disabilityterm·key differently based on different identities, such as age, sex or refugeeterm·key status, and thus possibly also the intersection of multiple identities. Depending on the measure used, certain demographic groups may thus disproportionately be categorised as disabled or not disabled; this could exacerbate exclusion if these measures are used to decide on resource allocation for individuals, health interventions or as disability statisticsterm·key to inform policy decisions.
We also see that the specific individuals who are categorised as disabled are not the same for each measure of disabilityterm·key. For example, while 8.8% of the population were categorised as disabled for both the WGSSterm and the 2QSterm measure, 4.7% were categorised as disabled using the WGSSterm but not the 2QSterm and 13.6% by the 2QSterm but not the WGSSterm. This aligns with earlier studies [ 17 , 25 ]. We conclude, as Beno Schraepen writes, that there is no one person with a disabilityterm·key but that different people have different characteristics that lead to exclusion or segregation [ 12 ] (p. 26). Each of the measures picks up people at greater risk of exclusion, but who the various measures categorise as being disabled deserves further attention.
Lastly, we see that, consequently, the conclusions drawn on the inclusionterm·dom and participationterm·key of persons with disabilities slightly vary based on the disabilityterm·key measure used. In this study, we used no schooling, being married/living together and having paid work as our indicators of participationterm·key. For all outcomes, we see that there are inequalities between people with and people without disabilities. However, the magnitude of that inequality varies based on the disabilityterm·key measure used, reflecting, again, that each of the measures captures a different group of people. This is important, as such disaggregated data inform decision makers in developing interventions, policies and programmes.
To understand whether society, policies, programmes and interventions are inclusive of persons with disabilities, we need to demonstrate how persons with disabilities are doing in relation to their peers. This requires comparing persons with and without disabilities using the same indicators. However, using different definitions of disabilityterm·key categorises a different group of individuals as disabled: although the WGE and the 2QSterm find a similar prevalenceterm·key of persons with disabilities in the population, the specific group that is categorised as disabled is not the same.
The findings demonstrate that there is no single, uniform category of “persons with disabilities” and that this has implications for subsequent outcomes on indicators of participationterm·key. Using different tools changes the extent of observed inequalities in key participationterm·key outcomes—in this case, education, paid work and partnership. This has significant implications, particularly when monitoring international frameworks such as the SDGs and the UNCRPDterm·dom, whose aim is to achieve equality for those with and without disabilities. We need to acknowledge that “different methodologies lead to a range of understandings of disabilityterm·key” [ 32 ] (p. 5118). Further research is needed to understand who the people are that different measures of disabilityterm·key are categorising as disabled and how this is of relevance for policies, programmes, (health) interventions and the monitoring of the UNCRPDterm·dom and SDGs.
Future research could delve into further examining how different measures of disabilityterm·key may impact the rights, social standing, participationterm·key, resource allocation and access to (disability-specific) resources of persons with disabilities. This involves conducting comparative studies using various disabilityterm·key measures to assess differences in access to legal protections, social services, and participationterm·key in societal activities.
It is important to recognise the diversity of disabilityterm·key and that the relationship between disabilityterm·key and participationterm·key will be shaped by intersecting characteristics and social identities. Data disaggregation by age, sex and type of impairment was not possible in our study due to limited sample size. However, a future analysis comparing measures should consider further disaggregation to provide a more in-depth understanding of these measures. Further research should also be conducted to analyse whether multiple measures of disabilityterm·key could be incorporated into data collection to capture a more comprehensive picture of the population that could be categorised as having a disabilityterm·key and are at risk of exclusion. This might be necessary to identify and address the needs of individuals who may be overlooked if only one measure of disabilityterm·key is used for categorisation, identification and resource allocation.
Limitations This study has several limitations. The measures of participationterm·key were restricted to schooling, marital status, and employment, which, while important, capture only a narrow subset of societal inclusionterm·dom. Further, although employment is an indicator of current participationterm·key, for education it is not known whether the onset of disabilityterm·key was before, during or after school age. Broader dimensions of participationterm·key, such as social connectedness, access to healthcare, and civic engagement, were not collected on the full study population and therefore could not be included in this analysis. This deserves attention in future research as these additional indicators could provide a more nuanced understanding of the barriers and opportunities experienced by persons with disabilities, particularly in contexts of displacement. The prevalenceterm·key of mental health conditions, including anxiety and depression, within the refugeeterm·key population studied may have influenced how individuals perceived and reported physical and sensory functioning. This could result in an overestimation of difficulties, particularly in a population already exposed to significant stressors and trauma (see [ 30 ] for more information). This study provides evidence that participationterm·key measures differ depending on the measure of disabilityterm·key used. It would be of interest to replicate this amongst a non-humanitarian population.
5. Conclusions The findings of this study underscore the importance of carefully selecting and clearly defining disabilityterm·key measures in both research and policy contexts, as it impacts the levels of participationterm·key measured when studying the inclusionterm·dom of persons with disabilities. Since definitions can both marginalise and empower individuals, it is crucial to understand and communicate the implications of the chosen measures. There is no “gold standard” to measure disabilityterm·key, but the most appropriate measure would be dependent on the intended purpose with eligibility for, for example, health screening programmes, requiring a different measure than monitoring the participationterm·key of those at risk of exclusion in social protection policies. Studies and statistics on disabilityterm·key should therefore always include the specific measure or definition used.
Acknowledgments We thank the study participants who generously gave their time to take part in the study. Thank you to the team at Mülteciler Derneği and Relief International who provided support to the project and team, and to the interviewers for their work, commitment and dedication to this study.
Author Contributions Conceptualisation, J.B. and S.P.; methodology, J.B. and S.P.; formal analysis, J.B.; data curation, S.P.; writing—original draft preparation, J.B.; writing—review and editing, S.P. and W.E.; visualisation, J.B.; supervision, A.S. All authors have read and agreed to the published version of the manuscript.
Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Istanbul Sehir University (ref 26/19, date 12 July 2019); Republic of Turkey Ministry of Interior: Directorate General of Migration Management (Ref: 72104824000/72304—date: 10 September 2019); London School of Hygiene and Tropical Medicine (Ref: 17623—date 20 August 2019).
Informed Consent Statement Informed consent was obtained from all subjects involved in the study.
Data Availability Statement Data cannot be made publicly available as consent foe this was not obtained from study participants. For more information about the data, please contact Sarah PolackpersName at sarah.polackpersName@lshtm.ac.uk.
Conflicts of Interest The authors declare no conflicts of interest.
Funding Statement This paper involved secondary data analysis of existing surveyterm·mth data and no funding was received for this analysis. The original surveyterm·mth was funded by a grant from the United States Department of State to Relief International. As part of this, Relief International awarded funding to the London School of Hygiene and Tropical Medicine (LSHTM) for research and technical support (LSHTM Grant Number ITCRZQ94). The opinions, findings and conclusions stated herein are those of the author[s] and do not necessarily reflect those of the United States Department of State or Relief International.
Footnotes References Associated Data Data Availability Statement Data cannot be made publicly available as consent foe this was not obtained from study participants. For more information about the data, please contact Sarah PolackpersName at sarah.polackpersName@lshtm.ac.uk.