Reopening schools: why, how and for who

Authors: Lucas Gortazar (World Bank) & Jorge Galindo (EsadeEcPol)

This is an updated version of the proposal that we published on 26 June, which includes latest epidemiological evidence. The objective is to ensure the safety of the educational community as much as possible by learning from new data available and experience in other countries.

From an educational and social perspective. After the decline of the first wave of the SARS-CoV-2 virus infection and the total shutdown of the economy during the months of March and April, a debate began on the need to partially reopen schools. The declared objective was to ensure the right to education, as well as preventing the widening of any educational gaps. Evidence from Catalonia shows that the response of pupils to remote learning has been very uneven.

Figure 1. Reasons why mothers do not help their children study: mothers with compulsory, post-compulsory or higher education

As Bonal and González (2020) show, during the first weeks of quarantine, almost 30% of pupils had practically no learning activity or connection with their teachers or tutors – and most of these pupils were from low-income homes. This is unsurprising given that family support for schooling, and access to a computer per person, are strongly related to the socio-economic level of families. These results are consistent with observations made in other countries, such as the United Kingdom and the Netherlands (Andrew et al., 2020; Bol, 2020).

Figure 2. Households with a computer per person, per income quintile

The right to education can therefore only be fully ensured by classroom or blended teaching. Consequently, quarantine has had a strong impact on equality of opportunity; it also lowers family expectations and perceptions of the rules of the game for educational progress. As Fernández Enguita (2020a) explains: "schools are responsible for the socialisation, care and education of children. In normal times we forget about socialisation, as we take it for granted in the existing conditions, but this comes to the fore when faced by a pandemic."

Quarantine has a striking effect on children's psychosocial development. Schools are not only a place of learning, but also the most important place for learning how to socialise. Absence from schools for long periods will have effects whose real significance we can only measure in the long term.

A failure to reopen schools can have a very considerable aggregate economic cost in the short and long term

From an economic perspective. Beyond the socio-educational aspects, we must consider the slowdown in the growth of human capital: the loss of learning caused by the closure of schools will have consequences on the training of young people, and therefore on the economy. An economic reopening is today the principal objective in terms of public policy, and the aim is to ensure that the economy recovers as much of its productive potential as possible.

Added to this is a more tangible effect on current labour capacities. There are eight million pupils under 18 in Spain today. A significant percentage of employees (and companies) may depend on a reopening of the classrooms to recover as much work capacity as possible for employees either working from home or onsite. Given that children under 12 need care, we are probably talking about:

1. full-time parents for children aged 0-6;
2. part-time parents for children aged 6-12;
3. a parent at home supervising children between 12 and 16, but able to work from home full-time.

A failure to reopen schools can have a very considerable aggregate economic cost in the short and long term. This cost will not be distributed symmetrically among the population. Therefore, whether for redistributive or aggregated reasons, sufficient funding and clear criteria for a safe reopening of schools is necessary.

Access to school is both a social right and an economic necessity, and so universality should be guaranteed as far as possible. However, epidemiological restrictions make a return to the old normality impossible and undesirable. The need to reduce the possibility of contagion (the mechanisms for which are detailed in the following section) requires the definition of two key elements when decision making:

1. A universal minimum of classroom attendance that will inevitably be below previous levels. This minimum will not be the same for each Spanish region, as regions vary in the incidence of the pandemic, and the physical size of schools tends to vary between regions (due to differences in population density). Minimums could range from 4 to 6 hours a day for all pupils. To guarantee this universal minimum, more teaching staff must be hired to reduce the size of groups.
2. Clear criteria on the prioritisation of pupils for extra classroom-based support that is focused on the wellbeing and educational and socio-emotional needs of pupils.

When public services cannot be delivered equally to everyone, the criteria for prioritising each group of beneficiaries usually reflects the political economy and social structure. In the case of the provision of schooling, many of the families that have difficulties balancing their job demands with childcare needs are middle-income earners.

When deciding who should spend more hours at school after reopening [1], a conflict arises between facilitating childcare for working parents (usually with secure permanent contracts), or helping children from poor families (with parents on temporary contracts).  We propose combining the two criteria as far as possible.

Priority criterion 1: socio-economic vulnerability

From an educational point of view, pupils from the poorest families are those who benefit most from tutoring and reinforcement in small groups or individually. This has an additional justification, as these children will have experienced quarantine in poverty with the worst physical and emotional conditions. Attention through personal tutoring (in small groups) is a strategy with a moderate cost and a strong positive impact – especially on pupils from poor families (Fryer, 2017; EEF, 2018).

Pupils from the poorest families are those who benefit most from tutoring and reinforcement

The United Kingdom has already announced a €1.2 billion programme that includes extra personalised modules for pupils who need attention (The Guardian, 2020).

To identify priorities, funding should be targeted at the reality of each pupil in each school, and this could be achieved using income tax or social security data (many regions already use such data when awarding scholarships and grants). However, in cases where this is not possible (as for example, in a recent Spanish government plan – PROA – to provide children with additional orientation and help), funding should be directly managed by the school according to the socio-economic characteristics of the pupils (reflecting, for example, the proportion of pupils from poor families). In these cases, schools will manage the prioritisation of the classroom-based tutorial support programmes.

Priority criterion 2: childcare

Low income partially overlaps with childcare needs. Higher or middle-income households are: (1) better able to afford childcare; (2) more likely to work from home or have flexible hours. However, this is not always the case.

When low income and an inability to work from home or on flexible hours are combined, the need for school support is greater. In the United Kingdom, for example, the children of employees providing essential services have had access to classroom education from the outset of quarantine.

When low income and an inability to work from home or on flexible hours are combined, the need for school support is greater

In the current phase, it is necessary to offer this option to parents who cannot work from home or with flexible hours. One possibility would be for employers to issue certificates confirming that an employee cannot work from home (always with the proper supervision of the corresponding inspectorate).

Schools can then request from the administration the additional funds needed for childcare and supervision personnel in properly equipped spaces. If such spaces do not exist in schools, then third-parties (mainly local governments) will have to offer municipal spaces.

Additional criteria: age and special needs

Normally, the need for extra educational care strongly correlates with age and specific educational support needs (including special needs, such as learning disabilities, or late entry to schooling).

Pupils with special needs require classroom-based support for more hours

Both younger children and special needs pupils require extra time in school. There is a natural age division between primary and secondary education, as children from 12 years onwards need less individual face to face attention. This is especially advantageous since we know that the probability of contagion in children under 12 is less that of children between 12 and 18 years. It makes sense, therefore, that the minimum classroom attendance of children in preschool and primary school should be greater, as well as the necessary investment.

At the same time, pupils with special needs, who are less independent in their learning (and who have less academic support at home), need classroom-based support for more hours. Often, however, such educational needs cannot be determined only by age, or whether the parents work from home, or income levels. Extra attention modules should still be activated in such cases.

Figure 3. Mechanisms of access to classroom teaching to ensure more equal opportunities and a reduced epidemiological risk

Who decides? Decentralisation vs. universality in criteria

The logical solution is that all schools provide a minimum essential classroom-based schooling for all pupils of between four and six hours (depending on location, infrastructure and resources). Additional modular sessions should depend on the four criteria described above.

To provide diversified and prioritised attention for the needs of each pupil, schools should remain open all day, rather than closing for the day at lunchtime. Currently, in some Spanish regions, most schools that serve vulnerable children close at 2 pm (Hóspido et al., 2019).

As Sintes (2019) points out, remaining open all day enables more individualised and flexible options for pupils, families and teachers. In the current situation, keeping schools open all day is especially important.

Reopening schools requires individualised and flexible models that reflect the reality of each school

But there are limits when applying the same criteria to everyone – and this will force difficult decisions to be made. Teachers best know the educational and family reality of each child, while employees and companies have detailed information on the possibility of working from home and with flexible timetables.

It is apparent that reopening schools requires individualised and flexible models that reflect the reality of each school. However, a lack of universal criteria could widen inequalities between schools and make technical decisions difficult. Therefore, it seems useful to provide obligatory general minimum requirements on a national or regional basis, while leaving sufficient margin for schools to achieve these minimums according to their circumstances.

Necessary resources

To ensure the minimum of classes and additional modules with the maximum of epidemiological safety, most schools will need more teaching staff and physical space.

For example: if, as we propose, pupils are divided into sub-groups, then teaching attention cannot be all day long for everyone. It seems logical that teaching staff and spaces are used to combine classroom teaching with independent online work (completed while at the school or in a municipal space). This requires the corresponding effort to fund more staff and more space, giving priority to those schools with the highest density of pupils per surface area and the highest proportion of pupils from poor homes (who need more face-to-face attention).

To this end, the €2 billion offered by the government is a good first step, although more funds will probably be needed. The regions must ensure that these funds are used for the priorities indicated, and that they mainly cover:

1. Staff costs and the modification of spaces for those schools with greatest needs based on the four criteria proposed: income levels (individualised tutoring and reinforcement); childcare needs (supervision and preparation of spaces); special needs support; and age (more teaching staff for preschool and primary pupils).
2. Catering services for schools without canteens.
3. Devices that enable more independent learning for pupils who do not require extra classroom hours and the prevention of potential closures of schools.
4. Economic incentives to ensure that absenteeism does not increase among the poorest segments of the population, through assistance and scholarships that are conditional on school attendance.

Following the model introduced with the new minimum income legislation that was approved in June (RD 20/2020, BOE), the actions carried out with this key investment should be independently evaluated so that adjustments, corrections and improvements in policy can be made.

There are still many uncertainties about the virus and its spread, but there are some points that seem clear (as long as we remain open to new scientific evidence).

• The younger the child, the less likely he or she will be infected. Data collected so far points to a lower rate of infection (Munro et al., 2020) for children below the age of puberty. However: (1) this rate is not zero, although the probability of infection is about half the community rate (and subject to considerable uncertainty); (2) this difference begins to narrow from the age of 12-13.
• Furthermore, very recent evidence from two separate studies in Chicago and northern Italy indicates that the viral concentration in those children with infection could be higher, possibly implying greater efficiency in transmissibility (see Haseltine, 2020).
• The spread of this coronavirus, like that of other closely related viruses, does not necessarily follow a linear pattern. Highly contagious events, in which one person infects many others, are responsible for a significant part of total infections. An analysis by the London School of Hygiene and Tropical Medicine, and the Alan Turing Institute, suggests that most infections are caused by a minority of cases.
• These highly contagious events are more likely when a number of conditions are met. The Japanese Ministry of Health defined such conditions as "the three Cs": close contact in crowded and enclosed spaces, in which saliva droplets that may contain the virus can easily travel from one person to another. It is important to note that evidence of contagion by aerosol (particles that remain suspended in the air) and not only by droplets (particles that are shot out) is increasing (Jiménez, 2020), which would help explain massive contagion events in closed spaces. Consequently, the use of masks, ventilation and the use of open spaces becomes as important as physical distance.

The sum of these indications poses a paradox for schools: although children are less susceptible to transmission, schools provide the necessary conditions for massive contagion events, especially as these spaces are not exclusively used by children (also teachers, administrative and support staff, and parents). Israel's recent failure to reopen schools serves as an example of the risk involved in this paradox: an excessively quick and careless reopening of schools triggered a surge in infections. Hagai Levine, professor of Epidemiology at the Hadassah School of Public Health (Hebrew University), stated in early August in the New York Times that the Israeli lesson for the world "is that you can open up the education system, but you have to do it gradually, with clear limits, and you have to do it very carefully" (Kershner and Belluck, 2020). Israel's mistakes illustrate how to do this.

The epidemiological basis for the return to school should be to minimise conditions favouring community or mass contagion events, first and foremost among adolescents and adults, and then among children. To this end, and always within the three framework models proposed above (full time, part time and mixed), and starting from the principle of respecting the independence of schools, there are at least four groups of measures that can be taken.

1. Contacts between adults. Define priority and inevitable contacts between adults, starting from previously established procedures that reduce meetings to a minimum and take advantage of online tools.
2. Biosafety conditions. Maintain conditions that reduce "the three Cs":
   • Adults and pupils over a certain age (possibly 8 or 10 years old, to be decided together with experts in child behaviour) must wear masks at all times.
   • Maintain continuous ventilation (as far as facilities and climate allow) in all enclosed spaces. When weather conditions and spaces allow it, a percentage of sessions (as many as possible) should take place outdoors. CO2 meters can also be used to estimate the degree of ventilation in classrooms and similar spaces. In the medium term, the advisability of investing in adequate air filtration and renewal mechanisms should be evaluated.
   • Ensure, as far as possible, distances of at least 2m between children and/or adults, and which may be less if they are exclusively between children under 12. When feasible, distances must be higher (up to 5m), or with separation mechanisms such as transparent screens.
   • When outdoor options are not available, organise spaces (such as classrooms, dining rooms, recreation, transport and meeting rooms) to minimise cross and frontal trajectories between people, and so reduce the risk of droplet dispersal.
   • Reduce rotations and shared surfaces. Assigning spaces, tools and furniture in a sustained way to the same groups of pupils.
3. Constant groups. Define small groups of constant contact, both inside and outside the classroom. Minimum group size will be determined by two factors: (a) the availability of staff; (b) age and ability of children to study independently. As stated above, it is a good idea to link spaces and people as much as possible, as long as adequate ventilation is maintained to ensure the biosafety of the space. This may be more complex in secondary schools, where groups rotate with teaching staff.
4. Stagger schedules. Within the margins that each situation allows, it is useful to stagger timetables to reduce density in the spaces used, hence the idea of a flexible "full-time school." This is crucial where the arrival and departure of pupils involves crowding in closed spaces.

It is relevant to note that the Israeli government tried to impose some of these conditions, but did not have the necessary resources to implement them. Even more, at a certain point, the government itself changed its criteria: a heat wave reached the country and it was decided to ease the previous requirement to use masks. Classes full of students without masks and with air conditioning recirculated pathogens. The lesson here is that when optimal epidemiological conditions cannot be maintained due to unexpected external causes, it is preferable to suspend classes for a few days, or to delay the start of the course until the necessary resources are available.

Another lesson from Israel is that care must be taken regarding contact between adults near schools. None of these measures, nor any that could be added, can be designed for all spaces. The independence of each region, and of each school, is essential to make the necessary adjustments for each physical and social context. But the objective remains the same: to minimise the probability of an outbreak in schools. However, it is necessary to have the tools to detect such an outbreak.

Community alert system

Over 60% of infected students in Israel were asymptomatic. To ensure the absence of contagion as much as possible, it would be advisable to carry out periodic random samplings at the neighbourhood or municipality level. This would help control whether the incidence of infections varies (something that, moreover, would serve for a whole series of public policy decisions at the local level).

The availability of diagnostic tests is, unfortunately, extremely limited. The epidemiological ideal (namely, a regular representative sample from neighbourhoods, districts and municipalities) does not seem feasible either in Spain or anywhere else. While this remains the case, it is necessary to establish a procedure that is accepted by parents and teachers. This should start by a report being made to the school of any symptom similar to a respiratory infection that is observed in a member of the school community, or any person who forms part of the family of a member of the school community.

Schools should nominate somebody to manage and organise procedures related to the pandemic

Ideally, the presence of such symptoms should trigger a diagnostic test for that person and their immediate contact circle, as well as a preventive quarantine while results are awaited. If a test is made and the result is positive, the school, or the school district, should decide how to proceed depending on the availability of diagnostic tests.

If tests are available, the collection of samples following the criterion of epidemiological clustering (risk group of infection) is advisable. If tests are unavailable, a total or partial preventive quarantine of all pupils and teachers in the school should be evaluated by the school according to the degree and intensity of contact of the positive case with the rest of the community. This method is not perfect, but it offers a relatively accessible warning procedure at the community level.

In operational terms, and as a final recommendation, schools should nominate somebody to manage and organise procedures related to the pandemic. An epidemiological professional from the public sphere, who understands the immediate context of the school, should also be nominated for consultations if needed.

In conclusion, it is worth remembering that we are dealing with an unusual phenomenon, and we must reconcile fundamental rights with recommendations from health authorities. The above proposals (in terms of how to open schools and the epidemiological considerations) are just the beginning of an inevitable learning and readjustment process.

To make these iterations as useful as possible, we should accumulate and systematically share knowledge. This will enable us to take advantage of the natural emergence of quasi-experiments to collect information on the effect of changing this or that parameter, the right to education and its prioritisation, and the needs of families and businesses. The results of such experimentation would otherwise be lost.

Notes

[1] For preschool pupils, given that there is little overlap between middle and low-income earners in the job market, the way in which points are given along these two axes will determine classroom access for children.

References

• Haseltine, WA. New evidence suggests young children spread Covid-19 more efficiently than adults, Forbes (2020)
  Kershner I & Belluck P. When Covid subsided, Israel reopened its schools. It didn't go well, New York Times (2020)
• Alison A, Cattan S, Costa-Dias M, Farquharson C, Kraftman L, Krutikova S, Phimister A & Sevilla A. Learning during the lockdown: real-time data on children’s experiences during home learning (2020)
• Bonal X & Gonzalez S. Desigualdades de aprendizaje en confinamiento. Resultados destacados (2020)
• Bol T. Inequality in homeschooling during the corona crisis in the Netherlands. First results from the LISS Panel (2020)
• Education Endowment Foundation. Small group tuition; Teaching & learning toolkit (2018)
• Fernández Enguita M. Un decálogo para la vuelta a la escuela (2020a)
• Fernández Enguita, M. ¿Cómo organizamos la vuelta a la escuela? The Conversation (2020b)
• Fryer Jr, R. G. The production of human capital in developed countries: Evidence from 196 randomised field experiments. In Handbook of economic field experiments (Vol. 2, pp. 95-322),  North-Holland (2017)
• The Guardian. English schools to get  £1bn to help pupils catch up after lockdown (2020)
• Hóspido L, Crespo L, Fernández M & Montalbán J. ¿Qué sabemos sobre el efecto del tipo de jornada escolar en el rendimiento académico? Dentro de "Indicadores comentados sobre el estado del sistema educativo español", de Fundación Europea Sociedad y Educación (2019)
• BOE. RD 20/2020.Real Decreto-ley 20/2020, de 29 de mayo, por el que se establece el ingreso mínimo vital (2020)
• Sintes E. Escola a temps complet. Cap a un model d’educació compartida, Fundació Jaume Bofill, Informes Breus #59  (2019)