Disability data for effective policy design: reflections from the TEACh project in Pakistan

By Nidhi Singal (University of Cambridge) and Rabea Malik (IDEAS, Pakistan), 26 February 2018

The lack of data on disability, as the United Nations Conventions on the Rights of Persons with Disabilities notes, has been the biggest obstacle to developing disability sensitive policies and programmes. However, the incorporation of reliable and simply worded questions on disability, as part of wider household surveys and national census, can provide invaluable insights regarding intersectionalities between disability and other factors, e.g., gender, in regards to access to education and socio-economic status. The data produced and insights gained can become the foundations for effective policy design.

Pakistan is one country where a focus on disability is still at a nascent stage. Both advocacy and policy making would benefit from reliable data, particularly on issues such as prevalence rates, quality of life experienced by persons with disabilities and the inclusion/exclusion of persons with disabilities from various mainstream processes.

Although Pakistan has historically collected information on disability prevalence, it has relied on unsophisticated tools to do so. The national census conducted in 1998 included a traditional binary question on disability: “Do you suffer from any disability?” Yes/No. Almost two decades later, the sixth national census undertaken in 2017, initially did not include a single question on disability. A last minute addition on disability was made to the existing questionnaire, in response to an intervention by the Supreme Court.[1] Nonetheless, the binary nature of the questions being asked and the process of adding these questions after data collection had begun is an interesting reminder of the need for the continued push for stronger mechanisms of holding institutions accountable for gathering disaggregated, robust and meaningful data on disability.

While specialized, sample-based surveys on disabilities are undertaken by international development organizations in Pakistan, for example a survey undertaken by the Poverty Alleviation Fund in 2012, their use remains very limited to stand alone programs. While some progress has been made in recent years,[2] having questions on disability included in multi-sectoral and large scale surveys, such as the National Nutrition Survey (usually covering 120,000 households across the country) has largely remained a struggle.

As part of the Teaching All Children Effectively (TEACh) project we used the Washington Group’s Child Functioning Module for children aged 5 to 17 years. We collected data across 1050 households in 3 districts of Punjab province. Our findings note some important results, particularly in relation to education, which was a key focus of the TEACh project.

Our survey suggests higher prevalence rates of disability than currently noted in Pakistan according to the previously mentioned surveys.  11.2% of children in our sample were reported as having moderate to severe difficulties in 12 (of the 13) functioning domains of the Child Functioning Module[3]. This is higher than other surveys in Pakistan, which usually note prevalence rates between 1 and 2%. One important reason for this difference, in our view, is how the Washington Group questions have been framed in a simple language which is easy to communicate and explain. Additionally, we undertook a very careful iterative translation process of these questions from English to Urdu. We trained our enumerators on how to administer the survey questions, but also helped them appreciate the underlying reasons for adopting a bio-psycho-social approach to understanding disability.

Our data suggests, that children who were reported to have moderate and severe disabilities were more likely to be living in relatively economically disadvantaged households. As shown in Figure 1.

Figure 1: Disability and household wealth

In the TEACh sample, the vast majority of 8-12 year olds without disability were in school, with only around 6% out of school. By comparison, around 23% of those identified with moderate to severe disabilities were out of school. Thus, severity of limitations in functioning had a significant impact on enrolment. This exclusion from schooling, we noted, was further compounded by the socio-economic status of the household. Enrolment rates for children reported as having moderate to severe disabilities in the poorest wealth bracket were 61.4% while for those in the richest wealth bracket were 87.5%.

Children reported as having moderate to severe disabilities were also less regular in attending school and more likely to repeat a school year. As shown in Figure 2.

Screen Shot 2018-02-26 at 11.43.17 am

Figure 2: Children with disabilities and school attendance and grade repetition

Notably there is also a pattern of higher absenteeism for children with no difficulties living in households which reported having at least one child with severe to moderate disabilities. It is important to emphasize here that these are also households which are poorer therefore multiple factors are likely to be shaping these patterns.

This small snapshot of a larger set of findings (which include issues around learning achievements) raises some important points.

  • The findings provide evidence on how children identified as having moderate to severe disabilities were noted as having differences in their school enrolments patterns, attendance and grade repetition.
  • Our data also suggests differences in school absenteeism in households where at least one child was identified as having moderate to severe disabilities.
  • Even though our sample size is small it is important to note that by situating disability questions in a broader comprehensive household survey enabled us to gather useful data on how disability intersects with a range of other variables, such as household income.
  • In the field, after a robust translation process and appropriate training, our enumerators found the questions in the Child Functioning Module to be simple and clearly worded. Parents were able to understand and respond to the questions without any confusion.
  • Insights such as those captured in our research, highlight the importance of gathering reliable disaggregated disability data for effective planning. With the availability of better approaches for framing questions on disability, such as the Washington Group set of questions, there is no reason to exclude disability as a category.

Dr Nidhi Singal is a Reader at the Faculty of Education, University of Cambridge. Her research focuses on children with disabilities attending a range of different educational arrangements, the quality of teaching and learning experienced, and the impact of schooling on short and long term outcomes in Southern contexts.


Dr. Rabea Malik is a Research Fellow at the Institute of Development and Economic Alternatives. Her substantive interests lie in the areas of sociology of education and political economy of education reform in low-income country contexts. Rabea has conducted policy research studies on the education-poverty nexus in low-income country contexts, the political economy of aid for development, and alternative service delivery mechanisms in education.


[1] A petition was put forward by a group of persons with disabilities to the Supreme Court of Pakistan. The court then directed Pakistan Bureau of Statistics (the federal department conducting the census) to include questions on disability. To comply, a set of codes were added to the question on gender to identify male, female and transgender individuals with disability. This was done after the data collection had started, and enumerators were already in the field.

[2] The Neilsen Household Survey on education covering 37000 households included questions on disability and the Multiple Cluster indicator survey carried out in Punjab, included questions on disability.

[3] While we administered the full set of child functioning questions, covering all the 13 domains, in our analysis we found a disproportionately high numbers under the domain ‘sad and worry’. This was noted by the Washington Group in other countries, as a consequence of which they adapted the ‘worry and sad’ questions to include the word ‘very’. For example, in the earlier version the question around worry was framed as: How often does (name) seem anxious, nervous or worried? Would you say: daily, weekly, and monthly a few times a year or never? Now the question reads: How often does (name) seem very anxious, nervous or worried? Would you say: daily, weekly, and monthly a few times a year or never? Thus, providing a more reliable picture of prevalence under this domain. Given in TEACh we used the earlier question, we have excluded this domain from our analysis.

Emma is the Washington Group Project Coordinator at the Leonard Cheshire Disability and Inclusive Development Centre at UCL.