Rural Ethiopian children, and members of their households, often suffer from common and preventable but debilitating illnesses, such as malaria, parasite infection and worms. Enrolment rates in Ethiopia are high, but school attendance is patchy, children often drop out of school (although they sometimes return), and grade repetition is common. This paper argues that the two phenomena are related: serious illness of children or in children's households is a major, and avoidable, barrier to children's schooling participation. This was demonstrated both in quantitative research, which analysed a 633-child longitudinal sample across 13 rural sites in Ethiopia, and qualitative research, a village case study, including interviews with 24 children and 10 caregivers.

In relation to child illness, the analysis found that health care was expensive and difficult to access. When children suffered from common illnesses, such as malaria, worms and diarrhoea, they were often absent for long periods. When they recovered, their schools often stated they had missed too much work to return to school, and they dropped out, but usually returned to school the following year. Quantitative analysis only demonstrates associations between variables, rather than causal relationships. It found that there was no association between illness and the probability that a child enrolled at all or on whether a child dropped out. However, illness increased the likelihood that children progressed slowly through school.

Analysis also considered children's time allocation to activities other than schooling. Sick children were less likely to do household chores, possibly because they were not strong enough, and were more likely to care for other household members, perhaps because they were at home and unable to do any other activities.

The paper also considers the effects of illness in children's households and found that children who had a sick adult in their households took on caring duties, chores and paid work. They were often late for or absent from school, but did not tend to drop out. If they dropped out, they often only "paused" schooling for a year and then returned to school afterwards. The effect of a household member's illness on children's schooling participation was more serious if children were involved in paid work to earn income for buying medicine and general household expenses.  Children often missed school to perform paid tasks or dropped out altogether to work.

The research found that parental death had similar effects to illness in the household on children's time allocation and schooling participation. In particular, children often "paused" school for a year after parents died. Quantitative research found that parental death did not affect children's time allocation, but this may be because of the way parental death was measured. The survey only measured if a death had occurred in the last four years, a relatively long time period, so the effects of parental death on children?s schooling participation may have passed by the time children were interviewed.

In relation to all three types of shocks, qualitative evidence suggested that inflexible teachers or school policies worsened the effects of illness on children's schooling participation. Some teachers were very proactive in attempting to keep sick children in school: they raised money for children or parents of children who were sick and needed to pay for treatment. But some teachers shut children out of class when they were late because of caring for sick parents. Others did not allow children who were absent for long periods (because of their own illness or caring for others) to return to school. Instead, teachers required children to wait until the next academic year, when they could resume school at the point in the year when they stopped attending. This could discourage them from returning to school altogether. If they were allowed to return to school in the same academic year, teachers rarely provided assistance, so children struggled to catch up the work they had missed. Children thus often had to repeat the year.

Quantitative data did not capture enough information to examine the effects of school responses to illness, although a recently collected round of data on schools will allow such analysis. The paper argues for further research into how schools can most effectively respond to illness. It also argues that local policy-makers and principals should urgently rethink how schools can be restructured to support children affected by illness.