Adolescents globally are vulnerable to mental disorders, injuries and unhealthy lifestyle choices, and this is heightened in young people who live in adversity. Early intervention programmes for adolescents can promote mental health, prevent mental disorders, and prevent risky behaviours and other unhealthy lifestyle choices. These programmes have the potential to make positive contributions to adolescents' health, and to their health in later life. Since about 80% of young adolescents globally are in school, intervention programmes delivered in the school setting could reach a large proportion of adolescents. We reviewed the evidence for school-based intervention programmes and found that most programmes come from high-income countries, and focus on one issue only, such as tobacco or alcohol use, or violence. Since many risk factors for poor mental and physical health, such as alcohol use and risky sexual behaviour, occur together, a programme targeting a range of factors could be more effective and more cost-effective. Such programmes may be more attractive to policymakers in low resourced settings. However, classroom-based programmes alone are unlikely to impact whole schools and sustain improved health benefits. Combining broader school climate and classroom-based programmes has been effective in reducing bullying and mental health symptoms in previous studies. The aim of this study is to address the evidence gap by developing and adapting a multilevel, gender-sensitive school intervention programme in low- and middle-income countries (LMICs) targeting a range of negative mental and physical health outcomes by promoting positive mental health; preventing mental disorders (specifically, depression and anxiety); and preventing a range of risk behaviours in young at-risk adolescents. We propose to develop the programme (Project HASHTAG) in two LMICs, South Africa and Nepal, co-producing two strategies: a broader school climate and classroom-based programme with adolescents, parents, teachers and school managers. For the first strategy, a school-level assessment, planning and action process, Thriving Environment in Schools (TES), will be conducted to improve school climate. School climate is defined as staff and students' perceptions of the school environment, which is influenced by the school organisational structure, relationships among staff and students, and their attitudes and behaviours. School climate improvement strategies support the role of classroom-based interventions by boosting commitment from teachers, staff and students, and by reinforcing content of classroom-based strategies. For the second strategy, adolescents ages 11-12 will participate in a group programme known as Thrive Together (TT), to equip them with skills identified in the evidence review as being effective in promoting mental health, and preventing mental health problems and risky behaviours. To develop and test these strategies, Project HASHTAG will comprise two phases. In the development phase we will interview stakeholders, and observe school environments and staff-student interactions to: (i) develop an outline of TES and (ii) gather suggestions for TT, with adolescents, parents, teachers, and other stakeholders in two schools per country. These suggestions will inform the work of an intervention development group in producing a TT programme draft. In the feasibility phase, we will assess the students before and after the TES and TT strategies are implemented in four schools per country using questionnaires. We will assess whether the questionnaires capture the necessary information on school climate, mental health and risky behaviours. We will also evaluate whether the processes needed to implement the programme work well. After programme completion, we will interview students and stakeholders about their opinions of the programme and whether it is feasible to implement in schools, and what will hinder or help this implementation.

The aim of the present study is to understand resilience and exceptionalism in high-functioning township and rural primary schools in South Africa. Previous research has shown that a large part of the explanation behind these schools' success is the leadership and management practices of teachers and particularly principals. Despite this near universal acceptance of the pivotal role of school leadership and management (SLM) for student achievement, accurate quantitative indicators of these practices remain elusive. Put simply, we do not currently have appropriate questionnaires that can accurately capture the school leadership and management practices among schools in challenging contexts in developing countries. One of the reasons for this is that these instruments are designed primarily with a developed-country-context in mind and do not account for possibilities that are prevalent in developing countries and typical in challenging contexts. Furthermore, in large-scale quantitative research, existing measures of SLM capture effective or ineffective SLM practices in superficial and fragmented ways. When looking at existing quantitative models of educational achievement researchers regularly find that there is a large unexplained component despite controlling for school resources and various student home-background factors. This is especially the case in challenging contexts where this disconnect between resources and results seems largest. One of the tentative explanations for this lack of explanatory power is that we are not currently capturing the true leadership and management practices (or lack thereof) in these schools and that this is partly due to inappropriate and inadequate SLM instruments. This is the first motivation for the inter-disciplinary nature of the proposed study; that the disciplines of Economics and Education bring different but complementary perspectives to bear on this issue of school leadership and management. Our previous research on schools in poor contexts in South Africa showed that deeper insights were obtained by a comparison between paired sets of schools with similar demographic and locational features, one performing poorly and the other performing strongly. This matched-pair approach is discussed briefly below. The proposed inter-disciplinary matched-pair analysis is, to the best of our knowledge, the first of its kind in either developing or developed countries. The current research uses 30 matched-pairs (matching 30 exceptional schools and 30 typical schools) because this provides the stark relief needed to identify which practices are driving the difference between the high performing schools and the average/low-performing schools in rural areas and townships in South Africa. The research will involve five stages: (1) Use population-wide assessment data to identify 30 exceptional primary schools (and their 30 matched pairs) in townships and rural areas in South Africa, (2) Conduct an in-depth study of 12 of the schools (6 exceptional and 6 matched typical) (3) Using the insights gained from Stage 2 develop new, more accurate and more context-specific measures of school leadership and management and pilot these in a different set of 18 schools (9 matched-pairs); (4) After finalising the new questionnaire this will be administered to all 60 schools to capture the SLM practices and behaviours of all matched pairs. In addition the team will administer background questionnaires to staff and students and monitor the Annual National Assessments in each of the 60 schools, (5) The final stage will involve validating the SLM measures identified in Stage 2, developed in Stage 3 and captured in Stage 4. The aim here is to use rigorous quantitative analysis to determine whether or not these new measures of SLM practices and behaviours are systematically related and specifically their predictive or explanatory power.

Context of research: To become sick with tuberculosis (TB), someone must first be exposed to someone who is coughing, become infected and then develop the disease. People with HIV and young children are more likely to develop TB disease once they are infected. One way to prevent TB is to find the people who live in a home with someone who has TB, check them and treat those with TB infection. This will prevent them from getting sick with TB disease. Many studies have shown that a drug called isoniazid (INH) reduces the risk of developing TB when given after being coughed on, so World Health Organization (WHO) advises giving INH to HIV-infected people and children under age 6 for six months when they have contact with someone with TB. Right now it is unclear what medicine we should give a child who has been exposed to someone with multidrug-resistant (MDR)-TB, when the germ is resistant to the most commonly used TB medicines, like INH. MDR-TB is becoming more common. The WHO estimated that there were more than half a million cases worldwide in 2012. Worldwide, it is estimated that at least a million children are exposed to MDR-TB every year. With new tests to diagnose TB quickly that can also detect resistance to the common TB medicines, the number of adults who are diagnosed with MDR-TB cases is increasing. In turn, the number of children exposed to MDR-TB is also increasing. Treating children who become sick with MDR-TB takes a long time (usually 18 months), usually needs a hospital stay, has medicines that may have many side effects, and is expensive. For these reasons, preventing MDR-TB in children is therefore very important. Until now, there have been no big and well designed studies to help us decide if using medicine to prevent a child with contact with someone with MDR-TB from becoming sick works. A few studies where doctors treated patients who have been in contact with MDR-TB have been done, but, each of them had problems. We think medicine to prevent MDR-TB might work, but a better type of study, a randomised control trial, needs to be done to prove that it works before we can be sure. Aims and objectives: We want to do a study in South Africa that looks at people living in the homes of someone with MDR TB disease. We will use a drug that doctors already use to treat MDR-TB called levofloxacin (LFX). We will test whether this medicine, given every day for 6 months, can prevent children from getting TB and/or dying. We will include children who live with someone with MDR-TB in the study. Children who get the medicine will be compared to those who get a sugar pill or placebo. This sugar pill looks like LFX but has no active medicine. Children will be followed for 24 months to make sure they do not get TB or have any side effects. We will also check if the medicine was easy to take, if it was safe, or if the TB became resistant to the LFX. We will also check how expensive it is to give this kind of medicine in the way that we think it should be given. Some children and their families will be asked to talk about their experience of the study and the medicine with the clinic staff. Potential applications and benefits: Until doctors learn what treatments work for preventing MDR-TB in children, it will hard to tell families what to do. If the medicine we are testing works to prevent MDR-TB, and is safe, and acceptable to families, we will be able to tell other doctors how to decrease MDR-TB all over the world. TB programmes will also benefit from this research because fewer children will get a disease which is costly to treat. Most importantly, if this medicine works, this study could greatly benefit children exposed to MDR-TB.

Learning to read for meaning is the most important skill that children learn in primary school. Yet children in 90% of South African schools are not acquiring this skill by the end of grade 4 (Spaull & Pretorius, 2019). The most recent Progress in International Reading Literacy Study (2016) (PIRLS-Literacy) indicates that almost one in every eight grade 4 students in South Africa cannot read for meaning in any language, despite the curriculum assuming that students can read in both their home language and English by the end of grade 4 (Howie, et al., 2017). Against this context, 'reading for meaning' among young children has recently been identified by South African President Cyril Ramaphosa as a top five national priority (South African Government, 2019). In adopting this new priority goal, South Africa needs to significantly advance the local body of knowledge on reading in African languages, identifying why children can't read, how reading could be improved, and how much improvement we could reasonably expect. In this study, we aim to fill some gaps in this body of knowledge. First, we will explore reasons for why children are failing to read with comprehension in African languages and English. Comprehension is what reading is all about - and this is what is measured in nationally representative surveys. However, reading comprehension is just the 'tip of the iceberg' with respect to underlying skills required to read (Spaull, et al., 2018). Necessary but not sufficient skills for comprehension include oral language, vocabulary knowledge, print awareness and decoding components such as phonemic awareness, letter-sound knowledge, word reading and oral reading fluency (Hoover & Gough, 1990). Using new emerging datasets on reading in African languages in South Africa, we will identify to what extent children can master these basics of reading across different languages. We will be able to construct a clearer picture of which of the reading 'building-blocks' children do and don't have, how these differ by language and whether reading skills systematically differ by poverty levels and gender. Fortunately these new datasets testing reading in African languages and English collect assessment data for the same children over time. With this data researchers can explore how reading skills develop and what gains in skills are acquired by initially low achievers, medium achievers and higher achievers. By observing best possible reading gains, it will be possible to get a better idea of the feasibility of attaining the presidential reading goal under current conditions or when conditions for improvement are created in schools. The constraints to learning at the school, teacher and classroom level are well understood in South Africa (see for example van der Berg, et al., 2016; Fiske & Ladd, 2004; Carnoy & Chisholm, 2012) with projects underway to address these constraints. But little is understood about underlying individual factors that may enhance or limit children's proficiencies in reading. Bullying, for example, is a very big concern in primary schools with South Africa recording some of the highest levels of bullying across all countries participating in PIRLS (Howie, et al., 2017). Bullying may reflect low underlying socio-emotional skills among children. Yet, international evidence and preliminary evidence from South Africa suggests that socio-emotional skills may be particularly important in fostering academic performance, including reading comprehension skills (Wills & Hofmeyr, 2018; Durlak, et al., 2011; Zins, et al., 2004). We will use available datasets to explore evidence on socio-emotional skills among primary school children in South Africa and identify whether indicators for socio-emotional skills (such as Duckworth's (2007) concept of 'grit') are linked to learning and reading skills in challenging contexts.

Many species, including humans, show morphological variation. Not all members of the species look the same, and they may vary in any number of traits. The presence of this morphological variation is very important, because it is a source of evolutionary change. Since populations contain individuals with different characteristics, natural selection may work to favour certain forms and repress others, or it may be the case that different forms are successful in different parts of the range of the organism, generating multiple species from a single starting species. In this project we aim to understand how morphological variation within a species is produced genetically, how variation is affected by natural selection resulting from ecological context, and how such variation is maintained when varieties meet. By understanding the production and maintenance of morphological variation we will gain greater insight into how evolution and speciation occur and provide input into models of how different species will respond to the various challenges that might occur as a result of climate change. We have developed a South African daisy species, Gorteria diffusa, as the best model system for this work. Gorteria produces classic "daisy" flowerheads composed of small circular disk florets in the centre and large elongated ray florets round the outside. All the florets are orange but some of the ray florets sometimes produce raised black spots that mimic the fly that pollinates the species. Within Gorteria's range in South Africa it exists as around 15 distinct forms (called morphotypes), each of which has a unique combination of floral traits, such as ray floret number and colour, presence or absence of spots, number of spots, presence and position of highlights in the spot, and presence and position of papillae in the spots. This species therefore provides an excellent example of extreme morphological variation, but is nonetheless easy to collect, grow and work with. The relative immobility of plants removes problems of migration and self-selection of environment. We have established procedures for molecular biological work with Gorteria. We can perturb gene expression using transgenic approaches, a very powerful way of understanding how genes control plant morphology. We have a good understanding of the molecular genetic basis of the development of a single petal spot type. We have also developed a strong collaboration with Dr Allan Ellis, a pollination ecologist at the University of Stellenbosch, South Africa, who will help with this project by providing support in the field. We have already defined how the different morphotypes of Gorteria are related to one another. In this project we will map different aspects of floral morphology and pollinator behaviour onto this phylogenetic tree to understand which direction evolution has taken for each trait and how many times each trait has evolved. Working together in the field, we will quantify the ecological context of each morphotype to understand how selection has favoured different morphologies in different geographic locations, and then test the hypotheses we generate by transplanting plants between different sites. We will then define the molecular evolution underpinning this morphological evolution - analysing what changes to key genes have allowed the visible changes we observe. Finally, to explain why morphologies don't merge into a continuum when populations meet, we will analyse the morphology and genetic structure of Gorteria populations at the places where different morphotypes meet, and explore post-zygotic isolation between morphotypes. Taken together, these data will give us an integrated eco-evo-devo understanding of how this enormous variation in flower types exists within Gorteria, providing us with insight into how species radiate so rapidly in the Cape Flora and other biodiversity hotspots.