The World Report on Disability (2011) cites a global disability prevalence rate of 15%, suggesting there are about 1 billion people in the world with a disability. However, censuses and surveys frequently come up with very different rates, often significantly lower. In the past this was usually due to faulty methodologies, using questions like “Do you have a disability?” that have been shown to seriously under-identify people with disabilities. The Washington Group on Disability Statistics’ website contains a lot of information about the inadequacy of these approaches, and how this led to the development of the Washington Group (WG) questions. However, even when using the Washington Group questions in a correct manner, countries tend to report prevalence rates lower than 15%, usually ranging from 6 to 12%. Sometimes this sparks concern. People have asked, “Are we missing people with disabilities? And if so, why?”
This question implies that there is actually some gold standard of what constitutes a disability and that the world is cleanly divided into people with and without disabilities, and we only need to ask the appropriate question to most accurately place people on either side of the line that differentiates between disabled and non-disabled people. In fact, functioning exists along a continuum. To identify a group of people with a disability, we need to find the most appropriate place on the continuum in which to place a threshold – where those above that place along the continuum have disabilities and those below it do not. The place selected (often called the cut point) should be selected to meet the needs for which the data are being collected. No single, absolute and universal standard exists that separates people with and without disabilities. Think about mobility. Some people have trouble rising from a seated position, others can do that but struggle to take a step, some can take a few steps, some can walk around a small room but no further, some can walk a kilometer, some can run a marathon. Where one draws the line that would determine whether a person has a disability depends on the reason for drawing the line. The important thing is that the reason is clear and suitable to the purpose of why the data are being collected. Different estimates of prevalence may all be valid in their own right even if they differ from each other as long as this purpose is well understood, and the methodology matches that purpose.
The need to define a point on a continuum to define groups of interest is not unique to disability. In fact, it is very common. Take age, for example. Age exists on a continuum but we often want to discuss the characteristics of children or of older people. How we define children or older persons depends on the use of the data. For some objectives older persons are defined as being older than 65, other objectives define older persons as over 85 years. In other cases, such as for some athletes, those over 35 are considered ‘old’. The prevalence of older age will vary depending on purpose for the determination and which cutoff is used and this is an excepted practice.
The WG questions define disability as those who have a lot of difficulty with (or cannot do) a core basic activity. The idea is to identify people who because of functional difficulties are at risk of being excluded from participation because of barriers in the environment. They were also designed to be internationally comparable, so that people with the same functional limitations living in different environments would most likely answer the same way. By combining the WG questions with outcome measures (like the SDGs) we can see the extent to which disability as defined by the social model of disability exists. Current estimates from around the world seem to fall in a fairly narrow range of 6 to 12 percent. This range could result from different age distributions in various countries, differences in other causes of disability, and different survival rates of people with disabilities.
The World Report (WR) on Disability’s figure of 15% was computed in a different manner, and that manner may lead to including people as having a disability who would answer “some difficulty” to one or more of the WG questions. This could still be an interesting population to look at. Maybe there is evidence that they also face barriers, but as a whole it is a difficult population to describe. Basically, the WR methodology creates an index of functional limitations based on aa large number of questions. An index is created that uses a statistical methodology (IRT) that makes assumptions about how the answers to various questions may be related. Once the index is created a decision has to be made as to where the place the cut point that divides the population into those with and without disability. The WRD looked at the scores on the index for those people with chronic medical conditions such arthritis, angina, asthma, diabetes, and depression –
and considered anyone with a similar functional limitation score or higher – regardless of having any medical condition – as having a disability. This manner of constructing the indicator is harder to explain to policymakers as its algorithms for combining large number of questions are not transparent. The indicator is also more difficult to compute, which is part of the reason why the WG methodology was designed the way it is.
Differences between prevalence rates from different methodologies is not unexpected or necessarily problematic – just be sure that you understand how disability was defined and when comparing prevalence rates, be sure you are comparing like methodologies designed for the same purpose. For SDG disaggregation and for international comparisons of disability, we feel the WG measure is the appropriate approach.