Many studies note the problem of the 'immigration paradox'. Immigrants have been associated with declining academic performance and aspiration. This line of research is cross-sectional and compares first, second and third generations. The Longitudinal Immigrant Student Adaptation Study (LISA) combines longitudinal, interdisciplinary, and comparative approaches to document adaptation patterns of 408 recently arrived immigrants from Central America, China, the Dominican Republic, Haiti and Mexico over the past five years. This article presents data showing patterns of academic performance. The Structural Comparison Model (SEM) provides insight into the factors that contribute to these patterns. The data suggests that supportive relationships are a significant mediator in the academic outcomes of young immigrants. Implications for future research are also discussed.

What is the cultural context of this article?

In recent years, the number of immigrant children in schools in the United States has risen sharply. These children experience challenges associated with migration to a new country, poverty, racism and discrimination, and violence at school and in society. The children growing up in poor neighbourhoods also have to deal with unemployment, violence, structural barriers and segregation due to race and poverty. Racism and ethnic discrimination have a negative impact on psychological functioning, self-confidence and physical health. Many immigrant children also go to schools that are less good and where, for example, there is too little staff and are too many students. Worrying about physical violence negatively affects the ability to learn and is bad for student-teacher relationships.

All of the factors above make an immigrant's adjustment more difficult, and they lower coping abilities. The children become more susceptible to academic failure. Although motivation to learn is better than that of the non-immigrant children, performance on tests, grade, attendance and drop-out rates are more negative. So despite the benefit of high motivation, optimism and more positive attitude towards school, immigrant children still perform worse than the non-immigrant children. However, some children are resistant, and do well.

There are four Protective Factors in the Life of an Immigrant Child:

• Academic self-effectiveness is the belief someone has in his/her competence to control his/her own learning. This factor partly predicts the extent to which a child will learn the new language and enter into new relationships.

• Proficiency in the English language is a predictor of academic success. It takes about 5-7 years to bring the language level of an immigrant to the same level as that of a non-immigrant child. High self-efficacy has a positive influence on the proficiency of the new language.

• Factors related to the family environment are very important. Two or more adults in a child's life increase the likelihood of financial resources, supervision, guidance and discipline. Multiple caregivers can lower social anxiety and support academic commitment and outcomes, among other things. There is a direct relationship between the child's education level and performance of the child, such as on tests and grades. Educated parents provide benefits such as extended vocabulary and computer access.

• Gender also influences school performances. On average, girls perform better than boys. Many factors can influence this. For examples, boys are more often the target of violence, physical aggression and racism. Immigrant girls also have more responsibilities at home, which can keep them off the streets. Boys have less meaningful relationships with teachers and perceive school environments as less supportive than girls.

Why is relational engagement important?

Relationships influence school performance, because the quality of relationships in the school setting is connected to successful academic adaptation. Social support at school is important for academic adaptation in both immigrant and non-immigrant children. Social relationships have protective factors, such as emotional support, cognitive guidance, and positive feedback. Relationships in school are, among other things, important for promoting socially competent behaviour within the classroom. Peer relationships support the development of psycho-social competences, such as providing support and relaxation in the face of anxiety. Peers can also help with homework and exchanging information. However, because most immigrant children attend poor schools where there is a lot of segregation, they have limited access to networks of peers who have a lot of knowledge.

Relationships with supportive adults in school, such as teachers and coaches, are important for the social and academic adaptation of adolescents, and especially for immigrants. Protective relationships with non-parent adults provide additional bonding, safe environments in which to learn cultural norms and customs, and information important to succeed in school.

What are academic and behavioural relationships?

An academic relationship refers to the extent to which a student is connected to what is discussed in class. This contributes to academic performance and includes cognitive, behavioural and emotional dimensions. The focus in this article lies on the behavioural part, namely participation around academic tasks, going to school, paying attention, behaving in class, doing homework and handing in assignments on time. There may be a gap between what is academically possible and what is achieved by the child.

What is the goal of this research?

The aim of this research was to investigate associations with the risk and protective factors mentioned above. A threatening school climate was expected to increase the risk of poor academic outcomes, because of the negative associations with relational and behavioural engagement and lagging language proficiency. High levels of self-efficacy and language development were expected to reduce this risk, because of the positive associations with behavioural and relational engagement. Relationships to school were expected to increase behavioural engagement to school and would be associated with academic achievement. The Structural Equation Model has been used to investigate mediation. Certain background characteristics were also expected to influence the outcomes, such as having at least two parental figures at home, being a girl and having better educated parents. These characteristics were associated with increased employment and better academic performance.

Which methods were used?

The study used data from the Longitudinal Immigrant School Adaptation Study (LISA).

• Recruitment: Schools in Boston and San Francisco with high numbers of immigrant students have been selected to participate. With the help of school staff, young people were identified who met the criteria (newly arriving immigrant with both parents from the same country).

• Interviews: Each year, interviews were conducted with the children in a language that the child preferred. The interviews lasted 1.5-2 hours and consisted of different types of questions, that were all asked verbally.

• Participants: (N = 407, of which 53 percent were female) of newly arriving immigrant school students from Boston and San Francisco. The age was between 9-14 years old at the start of the study.

• Educational level of the parents was on average 9.2 years with a range of 0-21 years.

• Income distribution: 70% of the households that participated were within the range of 0-40,000 dollars per year.

• Family structure: the family size was between 2-17 people. The Chinese lived in the smallest families and the Central American and Mexican children in the largest families. The majority of children had two parents .

• Characteristics of the area of ​​origin: most children came from rural (urban) areas.

• Reliability and validity: a protocol has been developed that is relevant for different cultural groups.

• Case studies: In the third year of the study, 75 students were selected for case studies.

Which measurements were used?

• Demographic data included parental education level, parental employment, and income.

• Proficiency in the English language was measured with the Bilingual Verbal Abilities Test.

• School violence was measured on a 10-point scale to measure the frequency of violence and bullying at school and in the neighbourhood.

• Academic self-efficacy was measured on a 7-point scale.

• Behavioural engagement was measured on a 4-point scale that focused on the behavioural portion of academic engagement.

• Relational engagement was measured on an 11-point scale and included statements such as' I can count on at least one adult at school.'

• Figures were the primary training dates.

What are the results?

In general, the academic performance of the immigrant children from the different countries is declining. The academic performance of children with Chinese parents remains roughly the same. The greatest decrease can be seen in the Dominican children. Most figures were stable for the first three years. Girls consistently outperformed boys over the five years, but the changes over time were the same, with both girls and boys decreasing performance. Behavioural engagement was lower in year five than in year three for every student, except for the Haitian children. The Structural Comparison Model was used for the difference between year three and year five, because language proficiency is an important indicator for academic achievement. The (average) behavioural engagement from years three and five is generally declining. Only children from Mexico show a slight increase. The largest decrease can be seen in children from the Dominican Republic and Haiti. Proficiency in the English language affects the average grade in both years three and five. Girls had a higher mark on average than boys.

The central element of the model used was the mediating role of relational and behavioural involvement between student background and academic achievement. Relational engagement was found to predict behavioural engagement in each year, and behavioural engagement predicted the mean mark in both year three and year five. Problems at school reported by the students were correlated with relational and academic engagement. School violence in year three predicted the mean grade in year five. Children with only one parent or children with low-educated parents had more problems at school. Girls and children with more educated parents had greater relationship involvement.

There was a negative relationship between self-efficacy at year three and the mean grade at year five. This was inconsistent with the general pattern and could be due to too much confidence at the start of the trajectory.

The average of the sample was 12 years. However, there were age differences between immigration groups. Children from Haiti and Mexico were younger, and this could potentially affect the results. However, an analysis in which the children were divided into two groups based on age showed the same relationships and patterns.

Two case studies

• Henry was typical of the immigration paradox. He was 12 years old when he immigrated from China with his mother. He started out with high expectations of himself and with great academic potential and showed a dramatic decline in academic performance over time. The big difference took place during the transition to secondary school, contact with the teachers diminished and he no longer dared to ask questions. He also felt isolated and went to classes less, his grades started to drop.

• Rosa had a large, supportive network of relationships with family members, peers and teachers. Her academic performance has remained stable over the years. She came to the US from Mexico as a 13-year-old girl. Rosa had two parents and six brothers and sisters. Rosa saw her two older sisters, who were studying, as an example and helped her younger brothers and sisters with homework herself.

What can be concluded from the results?

The current research was about the susceptibility of recently arrived immigrants from different countries to risk factors for low academic performance. In general, the immigrant children's performance declined between the third and fifth year of the study. The research highlighted the 'immigrant paradox', as most children started out with high motivation and ambition, but academic performance declined drastically over the years. There were, however, some differences between the countries. For example, Chinese children generally had the highest grades.

In general, relational engagement was an indicator of academic achievement. However, the predictive value of this factor differs, for example, between Mexican and Chinese children. It is important that more research is done on this.

Girls outperformed boys both at the outset and over time. Girls had better relationships with teachers, because of less conflict and closer relationships.

The Structural Comparison Model provided further insight into the general pattern. The prediction that a child's relational and academic involvement in school was mediated by background and academic involvement was confirmed. Relational and academic engagement mediated the negative effects of school violence on grades. It is important that children with more susceptibility to problems and low academic performance, such as immigrants, experience a warm and positive climate in schools. Unfortunately, this is usually not the case.

Academic self-efficacy was associated with higher grades in both years three and five. School-reported violence and threats in school were predicted to be negative for language proficiency, relational engagement and academic engagement at years three and five. Students with less educated and single-parent families reported more problems at school. What emerges is that many of the schools that the immigrant children attend have too few staff and too many students.

Consistent with previous research, it was found that English language skills were predictive of academic performance. The case studies show that various factors influence performance, such as having positive, social relationships, siblings, two parents and more educated parents. It must be taken into account that academic performance generally decreases somewhat as children become parents, but this decline is very strong among immigrant children.

What are the limitations of this research?

Randomization of the sample was not possible because children had to meet special inclusion criteria. This limits the ability to generalize the data. Still, the authors think the sample was representative of newly arrived immigrant youth. Subsequent research can also use immigrants from countries other than those used in this research. The use of culturally appropriate performance testing is also important. More research also needs to be done into resilience. This is because academic resilience does not necessarily mean that the child is resilient in all areas.

Current research has implications for future research and school policy. It is important to reduce racism and threats in school. Practices that enhance relationships are desirable.