Educators’ attitudes towards refugee pupils: intergroup contact and virtuous circles

ABSTRACT Educators’ attitudes are a key factor in the practices they enact, their expectations for pupils, and outcomes for pupils. For newly arrived refugee and asylum-seeking children, education is known to play a crucial role in the settling-in process. This study uses survey data (n = 295) and case studies of 17 educators at two schools in England to examine the attitudes of educators towards refugee pupils. It finds that educators had relatively positive attitudes towards refugee pupils and that previous relevant experience was associated with positivity. Educators at case study schools saw refugee pupils as assets, emphasised refugee pupils’ resilience, and blamed difficulties encountered on the education system rather than pupils. In the survey, previous experience teaching refugee pupils was associated with more positive attitudes towards them. While the direction of causality in this relationship was unclear, the case studies provided examples of educators attributing their positive attitudes to their interactions with refugee pupils and experiences working in multicultural schools. The study contributes to literature showing that intergroup contact improves attitudes towards outgroups and proposes that the relationship between school and individual educators’ attitudes could be seen as a ‘virtuous circle’ of positivity.


Background
Over 100 million people are displaced worldwide, including 32 million refugees and nearly 5 million asylum seekers (UNHCR, 2023).Around half of all refugees are children (UNHCR 2017), who are of school-going age and have the right to a quality education.The importance of school for refugee children is well-established -not only in terms of academic and language-learning opportunities but also in the promotion of socio-emotional well-being through daily routine, friendship, and fostering a sense of identity (Chase 2017;Fazel 2002;Ott and O'Higgins 2019).
CONTACT Caitlin M. Prentice caitlin.prentice@education.ox.ac.uk Moreover, individual teachers, teaching assistants, and other school staff have been identified as key players who are able to influence the educational experiences of young refugees (Aleghfeli and Hunt 2022;Baak 2019;Hek 2005).
Studies of educators and refugee pupils tend to focus on the practices that educators enact (in the UK, see for example Madziva and Thondhlana 2017;McIntyre and Abrams 2021;Peterson et al. 2017) or examine how national policies influence schools and education systems (Crul et al. 2019;Koehler and Schneider 2019;McIntyre and Hall 2018).In these studies, little attention is given to educators' affective states -such as how they think and feel about refugeeswhich includes their attitudes.The present study aims to address this gap by examining educators' attitudes towards refugee pupils and by exploring factors associated with more positive attitudes.By better understanding such factors, training and support opportunities can be more carefully conceptualised, targeted, and carried out, with the aim of improving educational experiences and outcomes for young refugees.

Attitudes and educators
It is widely recognised that educators' affective states play an important role in the actions they take, their expectations for pupils, and the outcomes for these pupils (Fang 1996;Nespor 1987;Pajares 1992;Pettit 2011;Yoon 2008).Furthermore, the effects of educators' expectations are larger for pupils from low socio-economic status backgrounds and pupils who have low prior achievement (Jussim and Harber 2005), which has relevance for many newly arrived refugee pupils.The concept of 'attitude' is difficult to define, however, and shifts between disciplines, authors, and across time.In her brief history of research on educators ' attitudes, Richardson (1996) notes that there was a peak in interest in the topic in the 1950s and 1960s, when researchers wanted to understand how educators helped or hindered the goal of racially integrated classrooms.This led to a number of studies examining educators' attitudes towards racial and cultural outgroups.Since then, research on educators' attitudes shifted somewhat to look instead at educators' beliefs -about subject matter, pedagogical strategies, and groups of pupils.
The boundary between attitudes and beliefs is fuzzy, however, and the two terms are used interchangeably in some of the literature.Nespor (1987) argues that beliefs, being cognitive, are actually a type of knowledge -but knowledge that is affective.Pettit (2011), on the other hand, contends that attitudes and knowledge are both types of beliefs -affective beliefs in the case of attitudes and true beliefs in the case of knowledge.Within psychology, one of the most common definitions of attitudes is the multicomponent 'ABC' (affective, behavioural, cognitive) model, which defines knowledge and behaviour as components of attitudes.In the present study, attitudes are defined as the affective state a person has in relation to an object -in this case refugee pupils -that can be detected through their behaviours and cognitions.

Attitudes towards refugee pupils
Relatively little is currently known about educators' attitudes towards refugee pupils.Among a small number of empirical studies, Kronick's (2013) paper 'Exploring primary school teachers' attitudes towards urban refugee education in Nairobi, Kenya' emphasises educators' practices rather than attitudes.More recently, Saglam and Ilksen-Kanbur (2017) and Ekin and Yetkin (2021) used surveys to examine Turkish educators' attitudes towards Syrian pupils, finding that mainstream teachers were 'eager' to engage refugee pupils (Ekin and Yetkin 2021) and that teachers with experience teaching refugee pupils were also more positive about their interactions with these pupils (Saglam and Ilksen-Kanbur 2017).
Given this limited literature specific to educators' attitudes towards refugees, the present study also looks to two related areas that can help provide a foundation for the research.First, a small body of literature measures people's attitudes towards refugees using attitudes 'thermometer' measures, where participants mark how positive or negative they feel on a scale of 0-10 or 0-100 (Louis et al. 2013;Schweitzer et al. 2005).These studies use undergraduate students as their samples rather than educators and have as their focus refugees or asylum seekers more generally, rather than refugee pupils specifically.Additionally, some studies of educators' interactions with refugee pupils mention affective aspects, even if attitudes are not the intended focus of the research.Studies of this kind have produced mixed results.For example, Lems (2019) observed an educator tell pupils in Switzerland that they were ungrateful for the opportunities provided through their education programme, appealing to negative societal narratives of asylum seekers being lazy and abusing social service provision.While this example clearly illustrates a negative attitude, it is important to note that deficit perspectives can stem from positive roots (e.g.compassion), but may still be negative if they emphasise vulnerability and result in a loss of power and autonomy for refugees (Ingamells and Westoby 2008;Rah et al. 2009).In a study of a specialised secondary school for new arrivals in Queensland, Australia, Keddie (2012) illustrated how negative attitudes can have a positive veneer.An educator at the school explained how newly arrived Rohingya girls -small in stature, dressed in saris, and perceived as 'cute' -were infantilised, seen as the 'exotic other', and thereby met with low academic expectations (p.207).
More recently, studies have documented educators' positive views of refugee pupils as reflected in their practices -with several of these studies set in England.McIntyre and Abrams (2021) describe a number of provisions for refugee pupils across the country, highlighting the ways in which educators support newly arrived pupils to feel safety, belonging, and success (Kohli 2011).While they focus on educators' practices rather than specifically addressing their attitudes, it is clear through the educators' actions that they recognised the strengths refugee pupils bring to the classroom as well as their needs.At 'Fern College', for example, educators made sure to give new arrivals thorough assessments, with a translator, so that pupils' abilities in subjects like maths were adequately measured despite their lack of English language skills.Similarly, Peterson et al.'s (2017) study of a comprehensive school in Kent describes how educators encouraged refugee pupils to become prefects and took a number of steps to ensure refugee pupils had access to tertiary education.These types of practices illustrate educators' positive attitudes through a recognition of refugee pupils' strengths.Finally, in Madziva & Thondhlana's, 2017 study of schools in Nottinghamshire, a resettlement caseworker commented on the variation in attitudes towards newly arrived pupils, noting that schools with more experience of refugee pupils had more positive attitudes towards receiving them: The attitudes have been different, depending on where.If schools have the experience of taking refugee children the attitude is completely different . . .there's a lot more leniency there and understanding that the children are going to struggle and will need extra support . . .but one school had a really bad attitude and they were saying 'what do we do with these children?' (p.951)

Factors associated with positive attitudes towards refugees
Although the literature on people's attitudes towards refugees is limited, the more general body of literature on attitudes towards immigrants and immigration is relevant.This literature crosses a variety of academic disciplines -political science, economics, psychology -and methodologies -experiments, small scale surveys, international polls.Across the studies, there is a highly robust and consistent finding that higher education levels are correlated with positive attitudes towards immigrants and immigration (see, for example, Dempster and Hargrave 2017;Hainmueller and Hopkins 2014;NatCen 2017;Wike et al. 2016).The evidence is strong for a link between education and positive attitudes towards immigrants, including refugees, however, the mechanisms behind this association are unclear.While the obvious assumption would be that more knowledge leads to improved attitudes, there is scant evidence linking these phenomena causally (Dempster and Hargrave 2017).Indeed, Ceobanu and Escandell (2010) have argued that other factors associated with education -such as higher socio-economic status and living in diverse urban settings -are as likely to account for the association as education itself.
Literature on intergroup contact theory also bears on the present study.According to the intergroup contact hypothesis (Allport 1954), contact between different groups should reduce prejudice between them, if certain conditions are met, including: 1) both parties having equal status, 2) both parties having common goals, 3) the presence of cooperation, and 4) the presence of support from an overseeing authority.Since Allport's seminal work, an entire field of study has emerged to test intergroup contact theory in a variety of contexts, from classrooms to neighbourhoods to Iraqi football pitches (Mousa 2020).A number of reviews and meta analyses have shown that while the details of the four conditions and their mediating pathways are uncertain, support for intergroup contact improving intergroup attitudes is generally robust (Paluck et al. 2019;Pettigrew et al. 2011).
Within the literature on intergroup contact theory, a selection of studies has specifically examined intergroup contact in the school environment.For example, Rissanen et al. (2015) found that pre-service educators in Finland who had experienced more interaction with diverse cultures were more positive in their attitudes towards Muslim pupils.Similarly, Glock et al. (2019) found that preservice educators held more negative implicit biases against ethnic minority pupils than in-service educators with experience in diverse schools, and that the pre-service educators were also more likely to associate cultural diversity with increased workload.Thus, a small number of studies find that educators with experience of pupil outgroups are more positive about those outgroups, concurring with the larger literature on intergroup contact.

Aims and hypotheses
The aim of the present study was to examine educators' attitudes towards refugee pupils and to explore factors that may be associated with more positive attitudes.First, the study describes ways in which educators showed positive attitudes towards refugee pupils at two case study schools.Second, two hypotheses are tested using survey data: 1) being more highly educated is associated with positive attitudes towards refugee pupils, and 2) having more experience with refugee pupils is associated with more positive attitudes towards refugee pupils.

Notes about terminology
The term 'refugee pupil' refers to any pupil who is a recently arrived refugee or asylum seeker, including unaccompanied asylum-seeking children (UASC).The term 'educator' includes class teachers, subject teachers, specialist teachers, teaching assistants and school leaders.
The term 'affective state' is defined as relating to feelings and attitudes.Within this, the term 'attitude' is defined as a consistent pattern of feelings, thoughts, and behaviours about an object -in this case, refugee pupils.'Positive attitudes', in turn, is defined as holding attitudes that are favourable -in contrast with negative, or non-favourable, attitudes.

Methods
This study forms part of a larger project examining educators' knowledge, attitudes, and practices with refugee pupils in one local authority in England.The study used mixed methods with two main components: 1) a survey of educators (n = 295) across the local authority and 2) case studies of 17 educators at two schools.Both quantitative and qualitative methods were selected to allow for an in-depth exploration of the topic.Through the survey, I aimed to find out what knowledge and attitudes educators had on a broad scale.Through the case studies, I sought to understand how knowledge and attitudes interacted with practices and why educators thought and acted in the way that they did.Data was collected concurrently, analysed separately, then mixed by theme.The research was approved by the Central University Research Ethics Committee at the University of Oxford (reference number ED-CIA-18-174).
Survey participants were recruited through their schools of employment.All headteachers in the local authority were contacted by phone or email to invite their participation.I attended staff meetings and briefings to introduce and distribute the questionnaire, which was distributed online and as paper copies.Items on the questionnaire included quantitative and qualitative measures of educators' knowledge about teaching refugee pupils and attitudes towards refugees and refugee pupils.Data for the present study comes primarily from a 'thermometer' measure of attitudes, where participants indicated how they felt about refugee pupils at their schools -or hypothetically, if none were present (with 0 being 'very unfavourable' and 10 being 'very favourable').This was followed by an item asking educators to explain their reasons for their thermometer score.Participants were also asked about their experience teaching refugee pupils, pupils with EAL, and pupils looked after by the children's social care system (Likert scale 1-4, from 'no experience' to 'extensive experience').Participants' furthest level of education achieved was recorded as: 1) some secondary school, 2) GCSEs or equivalent (exams at age 16), 3) A-levels or equivalent (exams at age 18), 4) undergraduate degree, and 5) postgraduate degree.
The case study schools were both located in the largest urban centre in the local authority.School A was a secondary school with a history of welcoming refugee pupils.Their provision for refugees was mainly run by the experienced English as an Additional Language (EAL) department and consisted of a range of programmes run in the EAL base and supporting pupils in mainstream classes.School B was a primary school with a history of educating pupils with EAL, who had received a recent influx of Syrian pupils on a government resettlement programme.At School B, all refugee pupils were immediately placed in mainstream classes.Within the case study schools, participant teachers were selected by school leaders; I approached relevant support staff to participate throughout my visits.I spent 7-8 weeks in each school as a participant observer and conducted an in-depth interview with each case study educator towards the end of the visit.My role along the participant -observer spectrum shifted depending on the age and stage of the pupils and the wishes of the class teachers.In younger classes and beginner EAL classes, I acted much like a volunteer teaching assistant, while in older classes I was more of an observer.
Quantitative data was analysed in SPSS version 27.Qualitative data was analysed in NVivo using a mix of deductive and inductive coding.For example, I set out with the overarching goal of looking for cases of educators' positive and negative attitudes but added more specific categories -eg.'support is good' and 'like any student' -as they arose.Following analysis, I brought the survey and case study data together by theme.Comparison between interview, observation, and survey data improved the robustness of the findings, both in a confirmatory manner and through 'initiation', or using contradictions to reframe questions (Greene 2007;Onwuegbuzie and Combs 2010).It is important to note that whilst this study mixes data from the survey and case studies, the survey sample and case study sample had key differences.Specifically, surveyed educators had a wide range of experience levels with refugee pupils, with many having no previous experience at all with refugees (n = 141, 49.5%) or only a little experience (n = 109, 38.2%); therefore, their responses to how they felt about refugees in their classrooms were often hypothetical.In contrast, case study educators had between 2 and 30+ years' experience teaching refugees, so their responses were based in direct experience.
Finally, given the sometimes politically charged nature of rhetoric around immigration, the question of social-desirability bias is particularly pertinent to this study.I took steps to minimise bias by making the survey anonymous and encouraging participants to be as honest as possible in their responses, even if they worried their views might be unpopular.Throughout case study data collection, I was open with participants about my background -being a primary school teacher with some experience teaching refugee pupils -and emphasised that my goal was not to make a judgement about them but to learn more about what they thought, felt, and did on a day-to-day basis.The results should be interpreted with a knowledge of both my positionality and efforts to put participants at ease -but with the acknowledgement that social desirability likely still played a role in data collection.

Survey: positivity and learning about other cultures
Participant educators were asked to appraise their feelings towards refugee pupils at their school -hypothetically if none were present -on a thermometer graphic with a 0-10 scale.A similar thermometer scale (0-100 instead of 0-10) has been used in previous studies to measure participants' attitudes towards various outgroups, including refugees, but this was the first time, to my knowledge, that it has been used to measure educators' attitudes towards pupils.In this study, educators' mean thermometer score for feelings towards refugees at their school was 8.62 (SD = 1.73).
Survey educators were also asked to explain the reasons for the thermometer score they gave.Respondents could select as many listed options as desired and select 'Other' to write in their own reasons.The option 'They help us learn about other cultures' was by far the most frequent choice, as illustrated in Figure 1, above.
In the open-ended response section for 'Other', comments such as 'it really contributes to the richness of the environment' and 'they've got languages and cultures that enrich the experience of students in the classroom' were common, as well as comments referencing the 'nice atmosphere' resulting from a multicultural school.Moreover, when the 87 responses for 'Other' were analysed thematically, nine had to do with refugee pupils being good for the school community by exposing it to diversity.'I believe that it is good for children to meet others who have had a different experience of life than their own (life experiences, culturally)', wrote one educator.'We live in a market town which is less culturally diverse than other areas of modern UK'.

Case studies: focusing on assets
Overall, case study findings complimented the attitude thermometer scores.Educators at School A and School B were mainly positive about refugee pupils.Like survey educators, case study educators frequently tied their positive attitudes to the benefits of multiculturalism.They saw refugee pupils as assets to diverse, multicultural school communities and saw multiculturalism as enjoyable and entertaining.At School B, nearly every member of staff and pupil I spoke with mentioned their 'International Evening' event.Once per year, school families originating from outside the UK were invited to bring in food, clothing, music, or other cultural practices to share.The annual nature of this event could be critiqued as an example of superficial, conveniently compartmentalised multiculturalism.However, it is also worth noting that according to School B educators and pupils, International Evening was one of the best-loved and best-attended events of the school year, enjoyed by 'international' and British families alike.Photographs on display in the school corridor portrayed a packed school hall full of colourful traditional clothing, tables heaped with food, and children and adults crowded around the presenters to see -and tastetheir offerings.
Educators also framed having diverse cultures at school as educational.They cited examples of pupils learning directly from their refugee peers -about countries, regions, and what it is like to experience war.At School B, class teacher Kasia 1 noted that the influx of pupils from Syria had inspired curiosity among her pupils and that they now knew much more about the country than if it had just been 'something they heard about on Newsround'.Similarly, at School A, EAL head, Margaret, related a story from years past when a geography class took their topic on the Rwandan genocide far more seriously due to their Rwandan classmate's presence, even though the individual never spoke about his experiences: The teacher was very scrupulous about not using him in the lessons, like look at him he's been through this or asking him questions directly.But the students knew where he was from and the teacher said he felt that was very instructive . . .Instead of just being a situation in a textbook, it was a situation that had brought that boy to their class.
On a more local level, educators relayed how encountering refugee pupils had helped other pupils learn to be more kind, caring, and helpful.They reported that upon arrival, some refugee pupils behaved in ways contrary to school norms.While this was sometimes framed as a strain on tight resources, several educators also framed it as a learning opportunity for the rest of the class.A class at School B, for example, received a new Syrian pupil mid-year who had never been away from his parents.His educator, Palesa, related how the new pupil's behaviour had intrigued another pupil, Jayden, who had a reputation for 'nightmare' behaviour: When Eisa came in, kicking and screaming and doing what Eisa was doing, Jayden was like, oh can I help him?He really just grew up that day and was like 'I need to help him'.And he reined himself in, that was the beginning almost of his change, having to see someone else so distressed.
In my observation sessions, I noted that a handful of other children in the same class had taken Eisa -who now entered the classroom with a smile -under their wing, in Palesa's words to 'basically mother him'.They frequently saved him a seat on the carpet and asked him to join their activities.Once, when Palesa ducked out the door to see if it was raining too much to go outside, Eisa became anxious and the girl next to him patted him on the shoulder and said, 'Don't worry, she's coming back'.
Educators also saw refugee pupils' multilingualism as an asset.Margaret argued that having a large proportion of new arrivals at School A benefited everyone because of the attention that had to be paid to language in general: 'We have such a mixture of people from so many different countries, refugee and non-refugee.And that's very positive because it makes language at the forefront in a school . . .and that's interesting for everybody'.Similarly, class teacher and EAL lead at School B, Safiya, said that 'it helps us to learn other languages.I'm not joking, it does'.Indeed, in my observation notes, I found that Safiya -who was herself multilingual -commented to her Year 2s on differences between English and other languages present in the class, encouraged children to share words translated into their home languages, and frequently told them how wonderful it was to know and learn multiple languages.Higher up the school, interviewed Year 5s spoke proudly with me about the languages they knew -once to the point that a monolingual English-speaker appeared to feel left out and announced that one of her parents was American, which was 'kind of another language'.
In addition to learning about different cultures and languages, educators also displayed positive attitudes towards refugee pupils through an emphasis on resilience and the related expectation that pupils do well at school.When speaking with me, they often presented refugee pupils in a positive light and emphasised the progress they had made rather than difficulties they encountered.At School A, Catherine noted that decades of teaching newly arrived pupils made her confident they would make progress, whether in a predictable linear fashion or a sudden 'quantum leap'.Similarly, Shauna, said that progress for new arrivals was 'more noticeable because often they come with very little English, so it's easier to track their progress because it's so huge'.At School B, multiple staff would sometimes describe the same 'success story' refugee pupils without me asking, for example over lunch in the staff room.They described a child in Year 5 who had only been in the country a year and a bit and was becoming one of the more creative, skilled writers in the school.Then there was the boy in Reception who had stopped crying at drop-off and the girl in Year 6, who had become smiley and sociable recently.In many cases, the staff relaying these anecdotes were not directly involved with the pupil in question; it was more as though they were passing along commonly accepted good news that reinforced their image as a welcoming and successful institution.This emphasis on the resilience of refugee pupils was also reflected in educators' practices.At both schools, for example, I observed educators call on pupils randomlyincluding newly arrived refugee pupils -showing their high expectations for all members of the class.
Educators also mentioned difficult or challenging aspects of having refugee pupils in their classes -but they tended to frame these as a problem of the education system rather than individual pupils.Several educators at School B noted that refugee pupils added to their workload and took time and attention away from others.Teaching assistant, Samantha, for example, said that: We've still got 29 other children maybe in our class that need our help and support and if we've got that one child who's really struggling, because they don't know the language, and they don't know phonics, they don't -it, it sounds awful but they take up a lot of our time, and then other children are missing out as well.
Similarly, Kasia noted that given the pressures on educators to promote progress for numerous other groups -pupils with SEN, 'quiet middles', 'high ables', 'white working class' -she felt that adding refugees increased her workload and stress levels.Yet she framed her concerns as complex and conflicting: Like everything these days in education, you love your -and you know my class nowand you love these kids to bits . . .And you know it's always these conflicting emotions.On an individual level you really love them and appreciate them and see them.But they make teaching life so much harder.Indeed, educators often pointed out that as much as refugee pupils sometimes added to their workload, so did many non-refugee pupils.Don noted that 'a greater chunk of your attention and time is drawn towards pupils who need more support . . .but that's true for all sorts of groups of students, not just those from a refugee [background] and it's not true of all pupils from a refugee background either.'Michael took the argument a step further, noting that while any strain on a system can have negative effects, the education system in Britain was already strained and blaming this on the arrival of refugees 'is just a scapegoat'.Finally, Kasia, who had initially referred to refugee pupils as a 'drain on resources', ended that section of the interview by reflecting that her negative feelings were a symptom of larger systemic problems: 'I think all of these things are literally a product of my -of teachers in England being overworked . . .we just spread ourselves so thinly'.

Experience matters: survey data
Scores from the attitudes thermometer were examined in the context of educator characteristics to see which characteristics were associated with more positive attitudes towards refugee pupils.Prior to data collection, it was hypothesised that educators who were more highly educated would have more positive attitudes towards refugee pupils and that educators with more experience teaching refugee pupils would have more positive attitudes towards refugee pupils.Related educator characteristics were also tested, including experience teaching pupils with EAL and experience teaching pupils looked after by social care.
In bivariate associations, educators' highest level of education completed was not significantly correlated with attitudes towards refugee pupils (r = 0.07).Experience teaching refugee pupils had a small positive association with attitudes towards refugee pupils (r = 0.19), but experience with pupils with EAL and looked after by social care were not significantly associated with attitudes towards refugee pupils (r = 0.12 and r = 0.02 respectively).See Table 1, above.
These results cannot determine the direction of causality.It is unclear, for example, whether it was educators' experiences with refugee pupils that shaped their attitudes or whether people who already had more positive attitudes towards refugees were more likely to select -and be selected for -jobs at schools where they interacted with refugees.However, data from the case studies can contribute to a deeper, more nuanced exploration of these processes.

Case studies: attributing positivity to contact
Data from the case studies suggests that educators' professional experiences and practices may, at least in some cases, have indeed contributed to the formation their attitudes towards refugee pupils, with school-level factors playing an important role in this process.Both School A and School B had policies and practices that promoted positive attitudes towards refugee pupils through the valuing of multiculturalism and diversity in general.For example, at School A, there were newspaper clippings in the reception area and corridors of refugee pupils in the news for their achievements, and a school leader's office shelves were full of copies of booklets of writing by pupils -in English and non-English home languages.At School B, during my visit whole school assemblies included presentations partially in Chinese for Chinese New Year, stories about Nelson Mandela and Malala Yusuf to illustrate various British Values, and a whole school sing-along in British sign language.In one notable example, Education level choices: 1) some secondary school, 2) GCSE or equivalent, 3) A-levels or equivalent, 4) undergraduate degree, 5) postgraduate degree.
the head teacher started off the Red Nose Day assembly -which normally has an amusing, costume-adorned jocularity about it -by speaking about the Christchurch, New Zealand mosque shooting that had just occurred and explicitly stating how much the school community valued its Muslim families.But were these school environment factors causal in terms of their relationship with educators' positive attitudes?Several case study educators directly related personal and professional experiences to their attitudes towards refugee pupils, and other outgroups more generally.At School B, Aaron said that he had been raised in a white, working-class community and often heard statements that he now recognised as racist and xenophobic, for example about immigrants from Somalia and Romania being criminals.Based on his experiences at School B, however, he said he now understood better why refugees came to England and was impressed by the recently arrived refugee pupils' progress at school.He added that seeing how many of the refugee families supported each other -for example looking after one another's children -had caused him to rethink his own cultural norms and wonder what British society could learn from immigrants.Many of Aaron's comments employed a 'deserving/undeserving immigrant' narrative related to which individuals worked and which were dependent on the state, a theme that was reflected in his framing of refugee pupils as hard-working.Nonetheless, he seemed to be in the midst of a substantial change in attitude that he attributed to his experiences of working at School B and interacting with refugee pupils.
Other case study educators commented on the shaping of their attitudes less directly, mentioning professional experiences related to how refugee pupils were doing in their classes and at their schools.Syed and Don both gave examples of watching newly arrived pupils settle into school, learn English, and excel academically, to back claims that refugee pupils often made quick progress.Similarly, the 'success stories' relayed to me by educators at School B were examples of how educators' experiences of refugee pupils were shaping their attitudes in a positive manner.These stories were often repeated around the staff room, meaning that individual educators' experiences and attitudes were amplified into school-level factors that in turn may have shaped the attitudes of other staff.The possibility of this type of 'virtuous circle' between individual and school attitudes will be discussed below.

Discussion
This study found that participant educators had mainly positive attitudes towards refugee pupils.Survey participants marked higher attitude thermometer scores, on average, than participants in previous studies.Case study educators emphasised refugee pupils' assets and resilience, and blamed difficulties on features of the education system rather than pupils themselves.This study also found that educators' positive attitudes towards refugee pupils were associated with previous experience teaching refugees and with a positive ethos and set of practices at their schools of employment.Below, I discuss how these results add to our knowledge about refugee education and consider implications for policy and practice.

Educators' attitudes: more assets than deficits
Previous literature alludes to educators' attitudes towards refugee pupils through studies of schools' practices with refugee pupils.Often these studies take the form of researchers arguing that educators engage in asset-based or deficit-based approaches with refugee pupils.Dabbous' (2019) study in Glasgow, for example, found that educators saw refugee pupils through a deficit lens due to their focus on EAL and lack of recognition of cultural differences, thereby 'inadvertently assimilating' them (p.224).Several commentators have argued that educators -and society more generally -overemphasise refugees' pre-migration trauma, thereby missing other important features of the individuals (Ingamells and Westoby 2008;Marlowe 2010;Matthews 2008).The present study, in contrast, found that educators had a mainly asset-based perspective, resonating with literature that describes examples of good practice with refugee pupils.Although these studies about practice do not make educators' attitudes the main subject of their research, findings regarding educators' attitudes are often commented upon as side notes or reflected in the practices enacted.Some examples include the commitment educators had to their refugee pupils in Madziva and Thondhlana (2017) and the 'empathy, understanding, and positivity' they expressed in Bailey (2011, p. 72).School case studies by McIntyre and Abrams (2021) and Peterson et al. (2017) document educators showing concern for how safe and welcome refugee pupils feel, and promoting positive social relationships and academic progress -much as in the present study.
While previous literature in this area tends to rely on interviews, the present study adds quantitative data through the attitudes thermometer measure.In the present study, a mean thermometer score of 8.62 was found.Previous studies have found lower thermometer scores.Louis et al.(2013), for example, found a mean of 6.86 (SD = 1.72) 2 across a sample of Canadian and Australian undergraduate students.Schweitzer et al. (2005) do not report mean thermometer scores but found that more than half (59.8%) of their sample marked negative 'temperatures' about asylum seekers -ie.below 50 on a 0-100 scale. 3Thus, by the thermometer measure, educators' attitudes were found to be more positive than in previous studies; however, differences in participant samples and survey subject may have accentuated the difference.
Previous studies asked about attitudes towards refugees or asylum seekers generally, rather than refugee pupils at an educators' school.Educators may have felt conceptually closer to their own pupils -even if hypothetical -than people out in society more generally.In a review of factors influencing people's attitudes towards refugees, Esses et al. (2017) identified the dehumanisation of refugees as linked to negative attitudes towards them.In this study, around half (n = 144, 50.3%) of the survey educators had at least some experience teaching refugee pupils, meaning that when they thought of 'refugee pupil', they may have been thinking of real individuals rather than an abstraction -thus humanising rather than dehumanising refugees.Louis et al. (2013) do not report whether their sample of students in Canada and Australia had regular contact with refugees, but it seems unlikely that it would be the case for at least half of them.
Furthermore, participants in the present study were all educators, in contrast with the undergraduate students of previous studies.Educators are, on average, well-educated and education is known to be associated with positive attitudes towards immigrants and immigration (Cowling et al. 2019;Dempster and Hargrave 2017;Hainmueller and Hopkins 2014;NatCen 2017;Wike et al. 2016).In this study, nearly 40% (n = 106, 38.1%) of participants had an undergraduate degree and nearly half of participants (n = 135, 48.6%) had both an undergraduate and a postgraduate degree, since qualified teachers in England are required to have an undergraduate degree in Education or a Postgraduate Certificate in Education in addition to a degree in another subject area, and the vast majority of respondents were qualified teachers (n = 250) rather than teaching assistants (n = 45).Another influential factor could be the perception of teaching as a 'caring profession', which could have implications both in terms of self-selection into the profession and in the development of educators' attitudes once part of the system.

Attitude formation: intergroup contact and questions of causation
A second aim of this study was to examine factors associated with educators' positive attitudes towards refugee pupils.In the survey data, it was hypothesised that higher education levels would be associated with positive attitudes; however, this was not found to be the case.One explanation for this result could be a lack of statistical power since only 34 out of 295 educators did not hold a tertiary degree.(These participants were more difficult to recruit due to their absence from staff meetings.)The result could also point to a 'buffering effect' (Fuochi et al. 2020), with less-educated participants feeling more positive than expected due to their professional remit as educators and to their exposure to different outgroups through their work.Further study of less-educated school staff's attitudes towards refugee pupils is required to test these hypotheses.
It was also hypothesised that previous experience with refugee pupils would be associated with more positive attitudes.This hypothesis was supported by the data; however, experience with pupils with EAL and pupils looked after by social care were not associated with positive attitudes.This finding concurs with Saglam and Ilksen-Kanbur (2017), who found that Turkish teachers with experience teaching Syrian refugee pupils were more positive about these pupils, and suggests that educators differentiate between refugee pupils, specifically, and pupils who share some of their characteristics.It also adds to a growing body of literature examining how intergroup contact theory plays out in a classroom setting.A handful of studies has found that educators with experience of pupils from different cultures are more positive about teaching such pupils (Glocket al. 2019;Rissanen et al. 2015), and there is wide empirical support for intergroup contact improving intergroup attitudes in a variety of contexts outside the classroom (Paluck et al. 2019;Pettigrew et al. 2011).Notably, however, in its original form the intergroup contact hypothesis states that four pre-conditions must be met (Allport 1954).Three of these conditions -both parties having common goals, the presence of cooperation, and the presence of support from an overseeing authority -are likely met in variable levels across different schools and classrooms.One condition, however, is decidedly not met in classroomsnamely that both parties, i.e. educators and pupils, have equal power and status.Thus, the results from classroom intergroup contact studies suggest that perhaps not all conditions theorised by Allport are necessary to create a positive effect.
A further consideration is that studies of intergroup contact in classroomsincluding the present study -show associations between contact and positivity but do not determine causality.While it is certainly likely that contact with pupils from an outgroup might improve educators' attitudes towards them, it is also possible that educators with pre-existing positive attitudes towards outgroups choose to work at schools where encountering these outgroups is more likely.While the present study did not set out to test the causal relationship between these factors, qualitative data from the case studies can offer some insights.First, as part of educator interviews, I asked participants how they ended up working at School A or School B. Responses did not refer to refugee pupils or newly arrived pupils with EAL more generally, with the exception of Margaret the EAL head.Rather, educators tended to give practical reasons, such as living nearby or being placed at the school as a student and then being offered a job.Moreover, several case study educators attributed their positivity to their experiences with refugee pupils, such as Adam's explanation of how he had come to recognise xenophobia in his upbringing.These facts suggest that contact with refugee pupils could cause educators to have more positive attitudes towards them -although the association is likely to be bidirectional, with pre-existing positive attitudes causing some educators to work at certain schools.

Virtuous circles: implications for policy and practice
The case study data presented here provide new insights into the relationship between individual educators and their schools when it comes to positive attitudes towards refugee pupils.In the cases of School A and School B, a link was observed between positivity in the school environments and the positivity of individual educators.Similar associations have been noted in literature dealing with educators' attitudes towards other subjects.Blake and Cutler (2003) for example, studied educators' attitudes towards the use of African American English in classrooms and found that there was an association between positive attitudes and working in schools that were supportive of multilingualism and multidilectism.Similarly, Hallam and Ireson (2003) found an association between educators' attitudes about ability grouping and the relevant practices at their schools.They made one key caveat, however: 'Whether this indicates that educators' views are influenced by their current working environment or whether they search out an environment which is conducive to their philosophy of education cannot be established from the current analysis ' (p. 354).Equally, in this study educators' positive attitudes towards refugee pupils may have been formed by the welcoming practices of their schools or educators with positive attitudes may have been attracted to work at such schools -or both.In the words of Hallam and Ireson (2003), 'it is likely that there are complex interactions between the two' (p.354).
In the present study, these 'complex interactions' could be seen as positive feedback loops or 'virtuous circles'.That is, the positivity around multiculturalism, multilingualism, and refugee pupils at School A and School B likely contributed to educators' positive attitudes towards refugee pupils.Schools, though, are composed of individuals.Individual educators organised events like International Evening, hung newspaper clippings about refugee pupils in the foyer, and selected assembly content that projected positive images of immigration.Given this circular nature of attitude formation, there should be multiple in-points to intervene to improve educators' and schools' attitudes towards refugee pupils.Educators have a solid foundation on which to build positive, asset-based attitudes regarding refugee pupils, and this could be capitalised upon through on-the-job support and training.Programmes of support could leverage intergroup contact theory, promoting conditions that encourage improved attitudes by helping educators and pupils to feel they are cooperating to achieve shared goals, and overtly displaying support (Allport 1954) through whole-school practices.
Finally, when addressing the targeting of training and support, it makes sense to briefly discuss the validity and utility of the concept of 'refugee pupils'.While refugee pupils are a highly heterogeneous group, they also have unique features that are worth considering in an education setting, such as the recent experience of forced migration and uncertainty around future immigration status, in addition to likely being additional language learners.Indeed, it has been argued that to ignore refugee pupils as a group -for example by not labelling or speaking about them as such -can have the consequence of 'invisablising' them and missing opportunities to track progress and provide appropriate support (Pinson and Arnot 2010).A previous study from this project found that educators recognised that there were substantial overlaps in strengths and needs between refugee pupils and other groups of pupils and identified a range of training and pedagogical knowledge as relevant to teaching the diverse individual refugee pupils in their classes (Prentice and Ott 2021).Thus, it is likely advantageous to recognise refugee pupils as a group, but training relevant to teaching refugee pupils need not be solely targeted as 'for refugees'.

Conclusions
This study found that participant educators had relatively positive attitudes towards refugee pupils.While the study was of selected educators and schools in one local authority, there are good theoretical reasons to believe that educators' attitudes beyond the study would also be more positive than the general population, due to their higher education levels and selection into a caring profession.Furthermore, while this study could not determine the direction of causality between educators' experience with refugee pupils and their positive attitudes towards them, there are also good theoretical reasons, such as intergroup contact theory, to believe that educators' positive attitudes were at least partially caused by their experiences.Thus, overall, the picture is optimistic regarding educators' attitudes towards refugee pupils, and there is potential to further develop this positivity -with the overarching aim of promoting positive educational experiences and outcomes for refugee pupils.

Figure 1 .
Figure 1.Reasons given for attitudes thermometer scores.Total is greater than 295 because participants could select multiple options.