Introduction

José, a 15-year-old ninth-grade newcomer from Honduras, described the feeling of being away from beloved family members:

Estar bastante tiempo lejos de nuestras familias, que a veces no podamos hablar con aquella persona que queremos. La experiencia de viajar a este país no fue tan fácil porque hay momentos, cuando vienes, hay momentos en los que no puedes dormir, hay momentos en los que no puedes sentirte muy cómodo porque hay cosas difíciles y cosas fáciles en este mundo que son muy diferentes a lo que yo he vivido.

[To stay a long time away from our families, sometimes we can’t talk to that person we love. The experience of traveling to this country was not so easy because there are moments when you come, there are times when you cannot sleep, there are times when you cannot feel very comfortable because there are difficult things and easy things in this world that are very different from what I have experienced.]

Newcomer adolescents spend considerable time in high school settings. Schools have an incredible opportunity to provide support for the whole child, which includes social-emotional, physical, psychological, and academic support, especially for adolescents who are newly arrived immigrants (Kreuzer, 2016). The term newcomers in this case study describes immigrant multilingual learners entering high school within the past three years and scoring between one and three English proficiency level as measured by the ACCESS for ELs assessment designed by the “World Class Instructional Design and Assessment” consortium or WIDA (WIDA, 2020). Newcomers in these three levels ranged from entering, emerging, and developing their English language skills when they first entered their new school; they aim to show growth in language acquisition throughout their educational experience (WIDA, 2020). This study occurred in 2019–2020 and focused on newly arrived high-school-aged students between the ages of 14 and 22, who were born in and spent two-thirds or more of their life in their home country. The 585 students enrolled in an International Academy, the site of this research, were the newcomers in this study, and the students in the study chose their own pseudonyms.

It is known that nearly 100,000 undocumented immigrant youth graduate from U.S. high schools each year (MPI, 2017); however, immigration patterns slowed “to a trickle” due to Trump administration immigration bans (Tavernise, 2019) and the colloquially-called Muslim ban (Gomez, 2018). Also, COVID-19 travel bans and border closures left migrants and their families stranded, separated, and vulnerable to the virus (Guadagno, 2020), and, therefore, delayed enrollment to schools. Newcomers arriving during their secondary school years face the greatest and most long-lasting consequences of their high school experience (Sugarman, 2017). A closer examination of belonging and support of newcomers is critical, especially during a time of what teachers called a hostile, anti-immigrant environment (Ee & Gándara, 2020). In a time when physical (social) distancing is the new norm, distancing has been a hurdle for immigrant students for many years, both physical (geographical) and metaphorical (the invisible line drawn by anti-immigrant rhetoric).

The focus on students entering at the high school level is intentional as research has shown that late-arriving immigrants, particularly those who first enter US schools at ages 12 and above, experience greater acculturative stress compared to younger immigrants (Mena et al., 1987; Patel et al., 2016; Portes & Rumbaut, 2006). This study did not focus on MLLs who entered the US in elementary or middle school, nor did this study focus on Generation 1.5, or second-generation students, or students who were born in the US to immigrant parents. Data from the US Census Bureau’s American Community Survey suggests that between 2000 and 2014, an average of 154,100 newcomer youth aged 12–21 immigrated to the US annually (Sugarman, 2017), and these MLLs are often positioned as English deficient or are even invisible (Mitchell, 2012). Systemic inequities and language-based discrimination against recent immigrants (García, 2009), who are primarily students of color, warrant further investigation into student belonging in this sociopolitical context. All newcomers, regardless of documentation status, are entitled to a free, public education until age 21 or 22 in most states (Sugarman, 2017).

The study focused on the experiences of high school MLLs being welcomed into the US education system. According to Mitchell (2012), “A multilingual learner is a student whose daily lived reality necessitates the negotiation of two or more languages” (p. 1), one of them being English. MLLs are at all stages of language development in both English and their home languages (Mitchell, 2012). Scholars argue that MLLs are more than a combination of two monolinguals (Brisk, 2006; Canagarajah, 2011; García, 2009). Brisk noted that MLLs are “influenced by a dynamic cross-cultural experience, rather than rigid cultural stereotypes,” and understanding this “is vital for designing school policy, classroom practices, and assessment procedures” (2006, p. 3). MLLs access their linguistic repertoires to translanguage (García, 2009) and draw upon their single united language system to codemesh (Canagarajah, 2011) when navigating classroom assignments, communicating thoughts and feelings, and building relationships. This dynamic co-constructed, sociocultural school experience is the center of the research.

Youth who immigrate during their teenage years contend and cope with three areas of transition: adolescence, migration, and entrance to high school. The literature overwhelmingly affirms that belonging can have long-term, positive impacts on academics, physical and psychological health, happiness, and well-being (Allen & Kern, 2017; Correa-Valez et al., 2010; Hombrados‐Mendieta et al., 2013; Khawaja et al., 2017; Kia-Keating & Ellis, 2007; Moallem, 2013; Walton & Cohen, 2011). Cultivating belonging is neither easy nor linear, yet it is particularly critical for immigrant students (Gonzalez & Padilla, 1997; Motti-Stefanidi & Masten, 2017; Suárez-Orozco et al., 2009). However, regarding feeling a sense of belonging to his school’s newcomer program, David, a 17-year-old tenth grader from Honduras, shared,

It’s like a family because they help you. They give you support; they make you feel that you’re important to them; they give you attention, respect, affection, and everything. So, yes, for me it’s like a family. We spend eight hours seeing each other, five days a week, so I do recognize that they are like my family.

David’s school is the site of this research; his school community is called the International Academy, nested within a larger traditional high school of more than 4000 students. The Academy is designed for high school students who had arrived in the US within the previous three years; students receive holistic and language development support on their journey to graduation from a US high school. This International Academy instructional model is part of a network of 30 schools called the Internationals Network for Public Schools, originating in New York City with current sites located across the United States. Historically, newcomers enrolled in these schools and academies have met graduation requirements within three to five years of entry to the US, depending on transfer credits from their home countries and other factors.

The body of literature on belonging has called for increased investigation among diverse student populations (Allen & Kern, 2017; Goodenow, 1993; Kia-Keating & Ellis, 2007; Schachner et al., 2018; Suárez-Orozco et al., 2009; Vera et al., 2018). Belonging describes an affective component of a student’s feelings of being accepted, valued, included, and encouraged by teachers and peers in the classroom and feeling like an important part of the daily activity of the class (Goodenow, 1993). A sense of belonging in students has been evaluated, researched, and captured qualitatively and quantitatively in the literature, but fostering it is not simple or straightforward (Allen & Kern, 2017). To explore belonging in this subset of a broader study, a large survey administration utilizing well-recognized Goodenow’s (1993) Psychological Sense of School Membership (PSSM) scale and semi-structured online interviews were analyzed to describe newcomers and their experiences in a large, urban high school. This study is unique as it utilized the PSSM with a newcomer population larger than had been previously studied and aimed to test and extend existing findings.

For immigrant adolescents, studies have found that the sociocultural factors that lead to a sense of belonging include feeling safe and secure to be themselves (Craggs & Kelly, 2018), feeling acceptance or rejection by their peer group (Moallem, 2013), building relationships with supportive bilingual staff and adults (Vera et al., 2018), having peers and teachers who were perceived to be interested in and respectful of newcomer development (Khawaja et al., 2017), celebrating multilingualism (Motti-Stefanidi & Masten, 2017), fostering a sense of identity (Perez et al., 2009), grouping students intentionally (Kessler et al., 2018), and providing social-emotional support (Kreuzer, 2016), among others. This study aimed to explore what newcomers identified as additional factors that contributed to belonging.

Further, the study’s population represents more linguistically and culturally diverse immigrant adolescents from a wider variety of countries—within Central and South America, Africa, and Central Asia—than previously studies have included. Few studies have researched adolescent MLLs’ sense of belonging in their new school community (e.g., Kia-Keating & Ellis, 2007; Schachner et al., 2018) during a period of heightened xenophobia (Ee & Gándara, 2020; Guadagno, 2020) and ongoing nation-wide examinations of systemic and racial injustices, such as the Black Lives Matter movement (see Terriquez & Milkman, 2021; Washington Post Staff, 2020). Additionally, building upon the construct of belonging, this study also extends and complements the research of Suárez-Orozco et al. (2008), who conducted a five-year longitudinal study of recently arrived students living in Boston and San Francisco regarding academic engagement, performance, and networks of relationships. Their study deeply analyzed academic, cognitive, relational, and behavioral engagement, and it bundled sense of belonging among many other factors within their relational engagement category. Their study and methods grounded this research and provided a foundation to build on newcomers' experiences with a sense of belonging.

This study explored what variables, if any, could predict belonging. Grounded in the broader belonging literature among youth, the hypotheses of this study were that female students would have a higher sense of belonging (Allen & Kern, 2017; Goodenow, 1993; Itzhaky & Levy, 2002; Sánchez et al., 2005; Smerdon, 2002; Wojtkiewicz & Donato, 1995), that a higher GPA would indicate a higher sense of belonging or connection to school (Close & Solberg, 2008; Roeser et al., 1996), that those who worked a job while attending school would have higher belonging scores (Suárez-Orozco et al., 2009), that belonging was related to strong peer and teacher relationships (Close & Solberg, 2008), and that belonging would increase over time in the program (see Schachner et al., 2018). To test these hypotheses, both quantitative and qualitative data collection methods were employed to provide descriptive data that one approach alone could not. This study informs implementable secondary school practices and state/district policies by adding newcomer belonging predictors and newcomer-identified school belonging components to the body of literature.

The sociocultural theories of Vygotsky (1978) underpin the study, which explores the variables, social structures, and experiences newcomers identified as being helpful to fostering belonging. Newcomers are adjusting to a new country, a new language, a new culture, and a new school (Suárez-Orozco et al., 2008), doing so with the support of others in a social environment. Consequently, a sense of belonging can affect newcomers’ personal and psychological well-being (Allen & Kern, 2017; Correa-Valez et al., 2010; Khawaja et al., 2017; Kia-Keating & Ellis, 2007; Walton & Cohen, 2011) as well as their academic achievement (Moallem, 2013). This research aims to describe quantitative survey results with qualitative interviews in a descriptive, single-site case study. The questions guiding this case study were as follows:

  1. 1.

    How do newcomers experience a sense of belonging in a secondary newcomer program?

  2. 2.

    What predictors, patterns, or trends exist, if any, for newcomers’ sense of belonging in a secondary newcomer program?

Study outcomes can inform secondary schools about what newcomers declare fosters belonging in their new school in a new country and language. This study aimed both to quantify and qualify belonging for a newcomer population that is continuously “overlooked and underserved” (Ruiz-de-Velasco et al., 2000, p. 1) by exploring in-depth demographics and belonging experiences in a social context.

Method

Descriptive case studies provide insight into complex issues and experiences; they describe a natural phenomenon within the context of the data that are being questioned (Lambert & Lambert, 2012; Zainal, 2007). The design was chosen to gain a rich description of newcomers’ backgrounds and belonging experiences in their new high school (Yin, 2018). Two data collection methods were used: a 391-participant survey and 14 semi-structured online interviews. Within well-designed case studies, quantitative methods help to define the case (Yin, 2018), and qualitative methods help to present an in-depth understanding of cases in which one data collection method is typically not enough (Creswell, 2013). Robust case study can result through leveraging the strengths of each method. In this study, a holistic, single-site case study was designed, utilizing both types of methods to achieve an in-depth understanding of the case, students enrolled in the International Academy, and their experiences with a specific phenomenon, belonging. Goodenow’s (1993) definition and measurement scale of a sense of belonging for adolescents was used to anchor the analysis. Additionally, interviews were utilized to enhance the survey results and were coded using emergent and a priori coding for Allen and Kern’s (2017) belonging framework.

Survey Methods

Survey methods provided a versatile and efficient way to collect perception data, demographic characteristics, attitudes, feelings, beliefs, and opinions from a large population. The research participants were enrolled at a single site and took the survey in class in a simultaneous, large-survey administration. The goal was to reach a census, or survey participation from all 585 students enrolled in the International Academy program. Data were analyzed quantitatively to find larger trends and patterns.

Site

To access hundreds of immigrant youth arriving within the past three years, the research site or “bounded case” (Yin, 2018) was a secondary newcomer program, called the International Academy, at a large, urban high school in Virginia. The research site is a member of the Internationals Network of Public Schools, which has been found to be effective in welcoming, educating, and supporting high school newcomers (e.g., Bajaj & Suresh, 2018; Jaffe-Walter, 2018; Kessler et al., 2018; Roc et al., 2019). The 30 International Academies and School sites provide instruction designed with language and content integration in project-based instruction with heterogeneous, collaborative student groups (Internationals Network for Public Schools, 2020). Additionally, a bilingual team of Spanish- and English-speaking counselors, social workers, administrative assistants, and administrators provides acculturation support.

Participants

This study gathered data solely from immigrant youth. The participant population was 585 newly arrived immigrant students. Participant inclusion criteria were as follows: (a) they had arrived in the United States within the previous three years upon enrollment, (b) they were between 14 and 22 years old, (c) their native language was not English, (d) they scored between one and three as measured by the WIDA assessment of English proficiency upon entry to US schools, (e) they may have had limited or interrupted formal education (SLIFE), and (f) they may have been refugees, asylum seekers, or undocumented persons. Persons having these characteristics constitute a rare population (de Leeuw et al., 2008). De Leeuw et al. (2008) state that in early adolescence (ages 12–16), cognitive functioning is already well developed, and it is possible to use questionnaires similar to those designed for adults. With this number of students experiencing migration as an adolescent in the same location, a carefully planned, group administration of the survey was used to capture as many participants as possible. The entire population of students enrolled in the program was targeted because all students have a unique history and world view.

The survey respondent sample was comprised of adolescents (n = 391; age 14–21; 45.2% female, 72.6% Spanish-speaking [37.9% Salvadorean, 17.4% Honduran, 14.9% Guatemalan]; 7.4% Arabic-speaking, 4.3% Amharic-speaking, and 2.5% Dari-speaking; see Table 1). Adolescents represented grade levels 9–12 and had a mean GPA of 2.53 (of 4.0); most had experienced a change in family structure when moving to the US (64%). A description of the adolescents participating in this study is critical to situate the findings.

Table 1 Summary of descriptive statistics of demographics of survey respondents
Full size table

Sampling Procedures

Total population sampling was utilized in this study. Of the target population of 585, all students consented and were invited to participate, and 391 took the survey. During the initial administration, 350 students participated, a response rate of approximately 60%, due to student absences and 13 students opting out. Absent students were invited the following day to take the survey should they choose, and an additional 41 students took the survey, bringing the response rate to 67% and a total of 391 participants, which surpassed the amount needed for a 95% confidence interval. A sample of more than 384 participants was needed for a 95% confidence interval for a population of 590 students (de Leeuw et al., 2008). The expected response rate for this survey was high due to its in-person, personalization, relevance, and advanced-notice qualities (Anseel et al., 2010). A uniform administration of a face-to-face survey helped to achieve a high response rate (de Leeuw et al., 2008).

The sample responding to the survey was representative of the total population in two of three categories (see Tables 2 and 3). The chi-square analyses revealed no significant differences for home country (χ2 [df = 3, N = 386] = 3.896, p = 0.272). There was also no significant difference from the population for grade level (χ2 [df = 3, N = 386] = 7.420, p = 0.060). Slightly more females responded to the survey than are represented within the population (χ2 [df = 1, N = 386] = 4.712, p = 0.032). However, Cohen’s w (0.109) indicates a small effect size, which is consistent with the research finding that females have a higher response rate on surveys than males (Smith, 2008). Therefore, the PSSM sense of belonging analysis controlled for gender.

Table 2 PSSM belonging score descriptive statistics
Full size table
Table 3 Chi-square goodness-of-fit results of sample in context
Full size table

Ethical Considerations

IRB approval was received, and informed consent opt-out forms and assent forms were obtained from the parents and adolescents, respectively. Survey forms were translated into Spanish, Amharic, Arabic, and Dari for participants and their families. Due to linguistic differences within this population, all directions and survey items were carefully chosen or worded in simplified English sentences, avoiding idioms, to ensure that translation into multiple languages would retain meaning and increase reliability. Back translation with interpreters familiar with the International Academy context and pilot testing with alumni were used. Back translation interpreted the survey back to English to ensure the translated text was faithful to the original source survey (de Leeuw et al., 2008).

Instrument

The survey instrument consisted of 73 questions designed to gather information from students in the following major areas: respondents’ backgrounds and demographics, opinions on their experiences in four primary structures embedded in their instructional program, and their sense of belonging.

School belonging was assessed by Goodenow’s (1993) Psychological Sense of School Membership Scale (PSSM), with 13 positively worded and 5 negatively worded items. This construct is the only scale on the survey that was summed and used as the dependent variable in all data analyses following all reliability measures. The PSSM includes items such as, “I feel like a real part of this school,” and “People here notice when I’m good at something.” Items on the PSSM contained a 5-point Likert-type format (1 = not at all true; 5 = completely true). These were then summed to produce a scale score ranging from 0 to 32.

There were 32 demographic questions and 19 open-ended or multiple-response questions to garner perceptions, opinions, and examples. Students were asked their opinions in the first half of the survey, and easier demographic questions followed to increase the likelihood that students complete the survey and avoid submitting incomplete data. Multiple demographic questions (32) enabled specific ways of aggregating students beyond home country, language, age, and gender. Research has supported the contribution of these variables as having a significant impact on student performance in school, their academic resilience, and their sense of belonging at school (Mena et al., 1987; Perez et al., 2009; Suárez-Orozco et al., 2008, 2009). Based on that research, the in-depth demographic indicators collected were the length of time in the program, age upon arrival in the US, self-report of grades, extracurricular involvement, personal valuing of school, family structure before and after emigration, family reunification, parental education and valuing of school for their children, number of siblings, hours worked per week, and rent obligations.

Surveying MLLs is a difficult task, and ensuring valid data collection despite linguistic differences was at the forefront. Suárez-Orozco et al. (2008) surveyed MLLs for a five-year longitudinal study on immigrant experiences in school. The content and wording of many items were inspired by their survey and interview protocols. Their permission is granted by their citation request here (see Suárez-Orozco et al., 2008).

Validity and Reliability

Many measures were taken to ensure the validity and reliability of this survey and its administration. First, acceptable internal consistency reliability was established (Cronbach’s alpha) for this sample using the PSSM eight-item scale and was 0.85 (greater than the generally recognized threshold of Cronbach’s alpha of 0.70). Internal consistency of this scale ranged from 0.73 to 0.95 across samples and countries (Abubakar et al., 2016; Goodenow, 1993; You et al., 2011) and was 0.80 in Goodenow’s (1993) initial study. Goodenow’s original survey (1993) was the blueprint for the response order for the present study.

There are many studies on the validity and reliability of the PSSM given in English (Goodenow, 1993; Ibañez et al., 2004; Sánchez et al., 2005), and there are a growing number of studies on the validity of the PSSM given in multiple languages. The PSSM has been widely used, mainly in English-speaking countries, and has been associated with increased competence and self-efficacy (Ibañez et al., 2004), increased school attendance (Sánchez et al., 2005), and higher grades (Booker, 2007). However, this survey was taken in the students’ preferred language; the largest groups preferred Spanish (65%), English (21%), and Arabic (5%). In alignment with this study, the PSSM scale has been administered in multiple languages (e.g., Cheung, 2004; Cheung & Hui, 2003; Gaete et al., 2016; Togari et al., 2011) and in multiple countries. To increase validity, the study of this scale is increasing internationally, such as in China (Cheung, 2004; Cheung & Hui, 2003), Japan (Togari et al., 2011), and multi-country settings (The Netherlands, Kenya, Indonesia, and Spain; Abubakar et al., 2016).

When using the Spanish PSSM scale specifically, Gaete et al. (2016) studied the use of the PSSM among Latin American adolescents and found that using only the positively worded questions in the Spanish version was valid and highly reliable. They studied the validity and reliability of this scale in a sample of 1250 early adolescents in Chile and found that the internal consistency of this new abbreviated version was 0.92. Because the survey in this study was taken in multiple languages, the majority being Spanish, Gaete et al. (2016) guidance was followed and the negatively worded items were removed. The translated negative items, cognitive development, and proficiency in reading comprehension for students may have caused difficulty in understanding the intercalated question format (Gaete et al., 2016); therefore, those questions in the scale were removed. In this study, the unidimensional PSSM scale of the positively worded items was used in English, Spanish, and multiple other languages to assess belongingness.

Survey Data Collection Procedures

Approval was received to invite all students enrolled in the International Academy to participate in the study for total population sampling. One week before the survey was administered, students took home a parental consent form, available in families’ preferred languages. The day before the study, teachers collected opt-out forms and student assent forms; provided explanations and reminders, such as to charge their Chromebook; and showed a short student-created instructional video of how students could translate the Google Form into their preferred language. The study took place in students’ first-period class on a Wednesday, as data showed higher student attendance occurred mid-week. The survey was administered with a teacher script to reduce potential bias; the survey was given in one session rather than dividing it, as multiple surveys tend to reduce response rates (Porter et al., 2004). No identifiable student information was collected. Students were provided with a personalized handout to record the correct GPA, teacher team, and WIDA score in the survey. Thirteen students who opted out or students whose parents had opted out were provided a handout with free-response questions about their experience in school and the ways they belong to reflect on the same constructs; their data was not used, but the opt-out survey recipients maintained a reflective class atmosphere.

Interviews

The second data collection method was student interviews. Student voice is a quintessential element of a sense of belonging, and, therefore, interviews were chosen to enhance what students wrote in the survey about their sense of belonging and school community experiences. Interviews can support survey data by suggesting explanations for the “hows” and “whys” of key events and provide insights reflecting participants’ perspectives (Yin, 2018, p. 118). The interview phenomenon of interest was the newcomers’ experience with dynamic connections with new people, a new school, a new language, and a new country. Because of COVID-19 restrictions, 14 participant interviews were conducted via Zoom, a video conferencing online platform familiar to students. All students chose their own pseudonym and its spelling. Nicholas et al. (2010) state that online interviews can create a non-threatening and comfortable environment that provides greater ease for participants when discussing sensitive issues. The interviews were to enhance, explain, or elaborate on the initial survey data to bring greater insight into the research question.

Stemming from Bronfenbrenner’s (1994) and Goodenow’s (1993) seminal work on sense of belonging, Allen and Kern (2017) devised a recent iteration of a sense of belonging framework, called the bio-psycho-socio-ecological model of school belonging (BPSEM), which shows that there is no single component leading to school belonging for an adolescent. Rather, belonging is a blend of biological, psychological, social, and ecological qualities and experiences. This framework was used to code the qualitative research data to analyze the complex layers of a student’s sense of belonging.

Participants

Purposive sampling (Chein, 1981; Merriam, 1998) was used to recruit interview participants. Using a criterion-based selection process, all students enrolled in the International Academy were sorted into three categories that indicated their academic engagement based on grades and attendance. Suárez-Orozco et al. noted that academic engagement is important for academic success: “Highly academically engaged students are actively involved in their education, completing tasks required to perform well in school” (2008, p. 42). Students were grouped into three categories: high- (0–2 blocks missed and grades higher than 75%), medium- (3–15 blocks missed and grades higher than 68%), and low-engaged students (16–43 missed blocks and grades lower than 50%). To be clear, these indicators of academic grades and attendance are not to be confused with a lack of desire or interest in school. The two criteria tools aligned with previous research as indicators of academic success and foreshadowed students’ dropping out of school (Rumberger, 2004; Suárez-Orozco et al., 2008). By no means is a low academically engaged participant uninterested in education; the participant may be very interested in graduating and learning a new language but may have other life circumstances thwarting that desire.

Within each category, recruitment began with a randomly generated list of students per category. Interview rounds continued to ensure representation in each group. Although equal proportions could not possibly be achieved unless every student was interviewed, loosely proportional representation can be seen with gender, grade level, home language, and age that represent the makeup of all participants within the high (5 students), medium (5 students), or low (4 students) categories (see Table 4). Interviews continued until data saturation. Overall, 42 students were contacted, and 14 agreed to participate with parental permission or with their permission and consent if they were over 18 (see Table 4).

Table 4 Summary of interview participant demographics
Full size table

Protocol

The interview protocol consisted of 27 questions, and the interviews lasted approximately 90 min. The questions were devised based on the survey analysis and students’ experience of belonging (e.g. “Can you tell me about a time when you felt like you were a part of the school community?” and “Can you tell me about a time when you did not feel a part of the school community?”) Hoppe et al. (1995) recommended that student feedback sessions last no longer than one to two hours to maximize students’ attentional resources. Probing questions were prepared, and frequently used, to follow up any responses that required more nuanced answers, such as, “Tell me more about that,” or “Why did you feel that way?” The interview was conducted in the language most comfortable for students. Interviews were conducted by the bilingual researcher/author (English/Spanish), who built a sustained relationship and level of trust with participants at their school. To mitigate bias, interview protocols received outside review, all consent/assent procedures were followed, students could opt-out at any time, and critical friends reviewed each stage of the interview process, including coding. The author acknowledges that while it does not eliminate all bias, the sustained relationship, extended commitment to students, and conscious reflexivity aimed to mitigate some notions of power and privilege as an honorary insider. Carling et al. (2014) state that honorary insiders can produce findings and access that other researcher roles and perspectives cannot produce. Interviews were transcribed to conduct coding and analysis.

Data Analysis

Surveys

The first round of statistical data analysis consisted of three parts, utilizing SPSS and NVivo. First, descriptive statistics regarding a wide range of student demographics were synthesized to provide context for the study population. Second, the eight items in the Sense of Belonging section were summed into a scale score to measure the larger construct of belongingness (Goodenow, 1993). Regression analysis was conducted between student demographics and the scale score of sense of belonging on the PSSM to describe the population’s sense of belonging at a snapshot in time. Cohen’s (1988) recommendations for large association (r2) of 0.25 or higher, medium association (r2) of 0.09 or higher, or low association (r2) of 0.01 or higher were used to determine which variables explained a statistically significant proportion of the variance in a sense of belonging. Alpha was set at 0.05. Five main assumptions for regression were both checked and met: independence of observations, normal distribution, homoscedasticity, linearity of data, and non-zero variances. Third, student open-ended response text data were coded for emergent codes and then the belonging framework (Allen & Kern, 2017) for patterns. The quantitative survey data are the focus of this manuscript, with qualitative data used for recommendations from students themselves.

Interviews

The student interviews were coded and analyzed iteratively in three rounds utilizing NVivo. There was frequent code-switching between languages during the interview, and these language switches were transcribed and retained in the language the students used for each comment. English interpretations were then added by a certified transcription and interpretation service. Utterances such as “uh,” “um,” and pauses were removed.

First, inductive coding was conducted for emergent themes in all the interviews. Inductive codes are developed by the researcher by examining the data (Creswell, 2013). Creswell (2013) recommends researchers be open to emergent codes, avoiding any “prefigured” codes that may serve to limit the analysis. By coding for emergent themes within the research question first, unbridled themes could develop. Next, codes were reviewed for secondary codes, to clarify their titles, add details, combine, and reorganize. Creswell and Báez (2020) recommend looking for “overlap and redundant” codes, then reducing the number of codes and collapsing them into five to seven themes that become the major headings (p. 162). This was an iterative process, while themes and categories surfaced. A mind map was created from emerging secondary themes (parent codes) with child codes/themes. Memoing and field notes were maintained throughout the process to provide a supplemental check on analysis and reflexivity.

The next step was the third round of a priori coding for the circular, BPSEM Sense of Belonging framework from Allen and Kern (2017). The coding of each interview began again with the framework saturating the mind of the researcher, identifying which layer students were referencing when discussing their experiences. Lastly, each layer of the framework was reviewed to look for common trends, patterns, and commonalities with the emergent coding. The results of the two rounds of coding helped to qualify the quantitative findings.

Results

Results were analyzed for statistical significance and implications for the student experience. Table 5 provides regression analyses conducted using students’ PSSM belonging score as the dependent variable to determine which independent variable or factor may have been a predictor of belonging. Notably, trends emerged to push belonging research in a new direction. For the open-ended survey and interview responses, findings consisted of four larger themes that described aspects of participants’ experiences that fostered belonging.

Table 5 Sense of belonging regression results
Full size table

Important Predictors

Of particular interest, significant results arose for the variables of gender, home language, and paying rent.

Gender

First, aligned to previous belonging research (see Allen & Kern, 2017; Goodenow, 1993; Sánchez et al., 2005; Smerdon, 2002), gender (being a female) explained a statistically significant proportion of variance in sense of belonging score, t(379) = 2.623, p = 0.001, r2 = 0.018. However, in this study, gender explained only 1.8% of the variance in a sense of belonging. Using Cohen’s (1988) power tables, this would be considered a low association. Studies have shown that Latina females are more likely to finish high school than Latino males (Wojtkiewicz & Donato, 1995), and they tend to have higher grade point averages (López et al., 2002). This finding raises questions about support for belonging for male students.

Home Language

Next, in this sample, home language did not explain a statistically significant proportion of variance in sense of belonging score, F(4, 375) = 0.285, p = 0.888. However, students who do not speak Spanish, Amharic, or Arabic (the three languages spoken most in the population) had a lower sense of belonging score on average. In this sample, the order of highest sense of belonging scores to lowest by language spoken was Amharic (MAmharic = 22.41, SD = 6.72), Arabic (MArabic = 22.34, SD = 6.33), Spanish (MSpanish = 21.73, SD = 6.44), all other languages, then Dari (MDari = 20.67, SD = 5.29). This finding raises questions for future research about belonging for students without many common-language peers, such as Dari speakers in this study.

Paying Rent

Third, 24.1% of students surveyed paid some of their rent, and 9.1% paid all their rent. Students who paid some or all their rent had a statistically significantly lower sense of belonging score. Paying rent explained a statistically significant proportion of the variance in sense of belonging scores, t(369) = −2.872, p = 0.004, r2 = 0.022. For those who were responsible for providing a roof over their head, paying rent influenced their sense of belonging.

Trends

Most variables did not have a statistically significant effect on their sense of belonging but did provide some interesting insights into trends (see Table 5), such as GPA, English proficiency level, grade level, time in the program, parental factors, and employment status. These variables alone did not explain variance in their sense of belonging and failed to show that they can predict belonging.

Grade Point Average (GPA)

This variable was hypothesized to have an impact on belonging, but the data showed otherwise. Interestingly, a student’s GPA was not a statistically significant predictor of sense of belonging scores, t(378) = 1.681, p = 0.094, nor was a student’s belonging score a predictor of a student’s GPA.

English Proficiency Level

As students’ English proficiency (WIDA level) increased, their belonging scores increased (see Fig. 1). However, students’ WIDA level did not explain a statistically significant proportion of the variance in sense of belonging scores, F(3, 376), p = 0.277. Interestingly, WIDA 4, 5 and 6 students had a lower sense of belonging score than WIDA 1. As English proficiency increased up to a WIDA 3, sense of belonging increased. Studies have related being able to express and understand the language with high academic performance over time (Suárez-Orozco et al., 2010). There was also a connection within student interview comments between a student’s grasp of English and their success in classes, performance on tests, and sense of belonging. Additionally, there is a phenomenon of decreased belonging at the end of high school that merits future research.

Fig. 1
figure 1

Belonging trends as grade level and English Proficiency score increase

Full size image

Grade Level

The evidence shows a connection between grade levels and language growth with a sense of belonging (see Fig. 1). As students are promoted to the next grade level, their belonging scores increase. For example, as students move from ninth to eleventh grade, their sense of belonging score increases slightly, but twelfth graders have the lowest sense of belonging score (M = 20.47; SD = 6.44). Importantly, when controlling for gender, grade level explained a statistically significant amount of the variance in sense of belonging, F(4, 375) = 2.946, p > 0.05, r2 = 0.030. However, despite its statistical significance, grade level did not explain much of the variance in belonging. As compared to a student’s first year in the program, a student’s sense of belonging score increases each year and then declines in their fourth year or more in the program.

Family Factors

This study explored many family variables as possible predictors of belonging in participants’ new school. Family separation and reunification were experienced by 64% of students surveyed, meaning almost two-thirds of the students experienced a change in who was responsible for them when they were growing up and whom they were living with now (see Table 6). Whether students were separated from a parent during the process of migration did not explain a statistically significant proportion of variance in sense of belonging score, t(361) = 0.699, p = 0.485. The study examined the level of parental educational attainment as a gauge of socioeconomic status (see Table 7), one factor that is particularly pertinent to the educational and academic attainment of their children (Suárez-Orozco et al., 2008). Although a mother’s level of education was not a significant predictor of sense of belonging score, F(3, 340) = 0.111, p = 0.953, students whose mothers finished elementary school or elementary and middle school had a lower sense of belonging score than those students whose mothers finished high school. Similarly, a father’s level of education was not a significant predictor of belonging, F(3,322) = 0.284, p = 0.837. Interestingly, students across all demographics whose fathers graduated from high school had the highest sense of belonging score (M = 22.28 out of 32). Regardless of levels of parental education, 86% of students said their parents think that getting good grades in school is very important. The study revealed that students place a very strong value on their education, and 98.7% of students feel that it is important to do well in school. This finding verifies the importance and value of education that students and families hold.

Table 6 Survey respondents change in family structure in the transition to the US
Full size table
Table 7 Parental education levels
Full size table

Working a Job

Lastly, students in this study were learning to balance working and going to school at the same time, and 40.9% of students were sending money back to families in their home countries due to the need to support family members and repay fees for moving to the US. Employed students did not have a statistically significantly higher or lower sense of belonging score. Importantly, 50% of the survey respondents had a job and were working and going to school at the same time (see Table 8). Perhaps most notably, 21.2% of students work 20 or more hours per week, all while attending school.

Table 8 Summary of financial responsibilities of newcomers
Full size table

Common Themes

Five themes regarding student belonging emerged in the qualitative comments through open-ended survey responses and interviews.

Language

First, the construct of language (all home languages and English) played a dynamic, sensitive, and nuanced role in developing a sense of belonging at school. This study went beyond the five components of language (i.e., phonology, morphology, syntax, semantics, and pragmatics; Berko Gleason, 2005) and focused on language as a social semiotic system, and sociocultural, receptive, and expressive method of communication (Halliday & Hasan, 1985). Using language gives way to socioculturally distinctive ways of thinking, acting, interacting, talking, and valuing (Gee, 1990). Students mentioned language frequently, what others said and how they said it, when describing themselves and how they felt they were accepted, valued, included, comforted, and encouraged by peers and staff. Specifically, translanguaging was used in the classroom daily and contributed to belonging. Students navigated assignments in two or more languages, accessing their linguistic repertoires to coach each other, communicate, and showcase their learning (García, 2009). Language differences, language acquisition, language learning, language challenges, and language fears wove through student comments. Language was also used to exclude and oppress, decreasing belonging. For example, Natasha, an 18-year-old eleventh grader from the Dominican Republic, felt excluded due to her Spanish accent, which differed from that of her Spanish-speaking peers from El Salvador. Language as a dynamic, social component contributed to newcomers’ belonging.

Network of Support

Second, this study revealed that newcomers built an intricate web of support from peers, staff, and families that contributed to their sense of belonging so that school felt like a second family or second home. The “IA is where students get to know and learn about a lot of cultures, languages, and the most important thing: they will have friends from other countries. Also, students can learn English and other languages easily” (Ben, a 17-year-old eleventh grader from the Ivory Coast). Ben’s observation highlights the network of international peers, friends, and staff who support their academic journey. Alexis, an 18-year-old twelfth grader from Guatemala, shared that students learn about themselves through experiences at school. The school “es como una familia, porque te ayudan, te dan apoyo, te hacen sentir que sos importante para ellos, te dan atención, respeto, cariño y todo. [School is like a family, because they help you, they give you support, they make you feel that you’re important to them, they give you attention, respect, affection, and everything.]” (David, a 17-year-old tenth grader from Honduras). When discussing how students provided support in L1, students shared that they explained, clarified confusion, took the initiative of helping others, called and texted peers to get them into virtual classes, and made friends while helping each other out.

Opportunity to Participate

Third, the opportunity to participate in activities, experiences, and practices in and out of high school fostered belonging. School structures that students referenced as opportunities were grouped into four categories: future preparation, tailored and extensive course offerings, during- and after-school activities, and resources. Experiencing these opportunities and activities together helped students belong and feel included, accepted, and valued.

Emotional, Linguistic, and Physical Safety

Fourth, emotional, linguistic, and physical safety and security were embedded within almost every interviewee’s comments. Feeling safe is a basic human need (Maslow, 1943). If a student does not feel safe, belonging cannot occur. Feeling emotionally safe to share your feelings, linguistically safe to make mistakes with language, and physically safe and secure from danger were foundational components of a rich school environment where newcomers feel like they belong.

Acknowledgment and Recognition

Lastly, students appreciated recognition of hard work, a celebration of successes, and feeling valued, seen, and heard. Goodenow’s (1993) definition directly addresses how feeling valued and encouraged by others defines school belonging. Acknowledgments of hard work along with a difficult life transition reinforced student value and showed that they are seen, heard, and appreciated. This recognition ranged from Honor Roll Assemblies and awards to personal words of advice and praise, to asset-based recognition and interaction with native speakers.

Many of the statistical results contradicted hypotheses about factors within a student’s life experience that may have affected their sense of belonging at school. Others shed light on the phenomenon of a student’s journey through their instructional program. Knowledge regarding the complexities of students’ transition into adolescence, a new language, a new country, a new school community, and, for 64%, a new family structure, provides insight into how school culture and programming may impact and support feelings of belonging, rather than variables that can predict belonging. These findings led to recommendations to foster belonging directly from students.

Limitations

Whereas many steps were taken to increase the validity and reliability of the data, there were a few limitations to each data collection component of this study. First, quantitatively, limitations of the survey methodology included social desirability bias in students’ responses and response rates for absent students the week of the survey. Although the sample size of 391 was quite large, low power resulted for many variables with few statistically significant results. Most of the variables studied could not predict belonging. For the qualitative data collection measures, 42 participants were contacted and 14 agreed to participate. Although the demographics of the students interviewed reflected a strong attempt at proportional representation, the student participants were not fully representative of the full population of the bounded case, which consisted mostly of Spanish speakers. Truthfulness in interviews and the survey and researcher reflexivity are limitations. The requirements for the protection of humans perhaps adversely affected the study because consenting to any kind of form may have dissuaded any particularly skeptical families. Lastly, common home countries, home language, age, or grade may not generalize to students from different places. This is always a limitation unless an interview were possible with all 585 students in the population. Although generalizability is not needed to draw important conclusions, this study’s findings, among other studies, can aid in generalizability.

Discussion

The central contributions to the body of literature is that this study reinforces existing knowledge and provides new insights on newcomer belonging.

Contributions

Regarding the study design, the methods used further confirm the validity of the unidimensional PSSM for multilingual youth. The large scale of multilingual adolescents taking the survey in 13 languages is unique and exemplifies the possibility and importance of careful methodological research planning including interpretation, back translation, pilot testing, and building on the work of others. The study occurred in 2019 and 2020 during a period of criminalization of immigrants, fear of deportation, and DACA uncertainties (Ee & Gándara, 2020; Ybarra, 2018) due to the harsh immigration enforcement policies implemented after the 2016 United States presidential election (Ee & Gándara, 2020). Belonging for newly arrived adolescents is impacted by these types of policies, sentiments, and rhetoric. Therefore, findings are relevant to current sociopolitical realities for immigrant youth and their sense of belonging to a school community. Future research could replicate these study methods at multiple types of sites such as Internationals sister schools, traditional high schools, and other secondary newcomer programs to compare and strengthen findings and aid in generalizability.

Regarding the quantitative findings, three of the 11 variables examined are found to be statistically significant predictors of belonging. The most significant results are that females had a higher sense of belonging; students from less-dominant language groups and students paying rent had the lowest sense of belonging; and as students progressed from ninth, tenth, then into eleventh grade, their sense of belonging increased until their senior year (see Fig. 1), making only two of the hypotheses true for this population. The statistical analysis failed to show that many variables were predictors of belonging.

This type of statistical analysis has not been conducted on such a large and diverse sample of newcomers and warrants further research. The results align with previous research regarding female students and the importance of relationships; results extend belonging research to include particular attention to students in smaller language groups and students paying some or all of their rent. Furthermore, the findings extend Suárez-Orozco and colleagues’ (2008) study to add that experiences, gender, home language, and paying rent may decrease belonging, and that some hypothesized predictors cannot predict belonging, such as GPA, working a job, or family factors.

Additionally, patterns were observed for students’ grade level and English proficiency. Future research is also warranted into trends in English language development and belonging (see Fig. 1). The phenomenon of belonging decreasing during senior year and as students attain higher English proficiency levels is an additional area for future research. The evidence suggests that there is a connection between grade levels and language growth with a sense of belonging (see Fig. 1). The trend results push belonging research in a new direction. The evidence suggests that many students connect language growth and acquisition with an increased sense of belonging. The longer students were in the program, the more English they acquired, and the more they felt they belonged. As students are promoted to the next grade level, their belonging scores increase. As students’ English proficiency (WIDA level) increases, their belonging scores increase. Students enter the program between a WIDA level one and three and maintain their enrollment in the IA as they aim to increase to a level six English proficiency—six is the highest score, called “reaching” native-English proficiency (WIDA, 2020). Studies have related being able to express and understand the new language with high academic performance over time (see Suárez-Orozco et al., 2010). Students’ comments show a relationship between their grasp of English and their success in classes, on tests, and belonging.

When coupled with other studies, schools may consider adjustments in their programming to support students based on this data, specifically within the populations of male students, students of smaller home language groups, and students paying rent.

Regarding the qualitative findings, evidence from this study describes how students are experiencing belonging in school. Five themes emerged as the components of students’ school experiences that contributed to their sense of belonging. The themes of language used to orient and encourage, networks of peer and teacher support, opportunities to participate in activities during and after school, safety (linguistic, emotional, and physical), and acknowledgment and recognition of student growth were critical contributing factors to newcomers’ sense of belonging. Stemming directly from these themes and student comments, district and school leadership may reference the following student-derived recommendations.

Considerations

Based on these findings, educators, researchers, and advocates may consider what students shared as recommendations to support belonging. These recommendations either came directly from the participants or emerged through the analysis of the open-ended survey responses and interviews. For male students and students who are members of smaller language groups, schools may consider the following five recommendations to foster belonging in secondary school spaces.

Create Peer Networks of Support

Support networks for peers who have a shared immigration experience serve a specific purpose: students feel that their classmates and friends play distinct and important roles in their acculturation, comfort, and experience at a new school. Recommendations include implementing ways to enhance peer relationships in and outside of the classroom, implementing collaboration in heterogeneous groups into classroom curriculums, and creating formalized peer mentorship programs. Many students mentioned that they love when they share ideas, hear from each other, and participate in Community Circles. Sharing ideas can also occur beyond the classroom walls in shared text groups (WhatsApp), and informal spaces such as hallways, cafeterias, and buses. The strength of peer relationships or bonds cannot be understated. Within peer groups of four, students explained that they helped each other navigate assignments. For example, David shared why he prefers to collaborate in small groups with others:

Por ejemplo, si no le entiendo a algo y otro alumno sí y está en mi grupo él “me echa la mano,” me ayuda a entender; ayuda bastante eso de los grupos de cuatro porque si cae uno están los otros tres que lo pueden levantar, entonces es útil. [If I don’t understand something and another student does, and he’s in my group, he “gives me a hand.” He helps me understand. It helps a lot to work in groups of four because if one falls, there are the other three who can lift him, then it’s useful.]

These planned, collaborative, classroom structures provided opportunities for students to belong. Although many students informally mentored new students as they arrived through mixed ninth- and tenth-grade classes, more formal mentorship programs can provide structured opportunities for all students to mentor and be mentored. These relationships provide social-emotional and academic learning support as a part of a newcomer transition.

Safety

In this study’s open-ended questions and interview responses, emotional, linguistic, and physical safety were critical aspects that contributed to newcomers’ belonging. First, emotional safety, which consists of confidence, trust in others, and freedom from embarrassment, is imperative for belonging. Next, linguistic safety to take risks with language and the freedom to make mistakes are critical to belonging. Lastly, physical safety (freedom from danger and violence) is one of Maslow’s (1943) basic needs and is also integral to belonging. Consistent with previous research (see de Graauw, 2014; López-Bech & Zúñiga, 2017) students recommended creating a culture of trust, respect, and patience, and providing school ID cards. Students frequently mentioned that they had a profound respect for classmates and friends, their points of view, and their learning. Students also referenced their teachers’ patience while “asking a million times” (José) as a critical quality of peers and staff when working together. Through trust, respect, and patience, students felt safe sharing feelings and ideas and, therefore, feeling emotionally and linguistically safe. For example, López-Bech and Zúñiga (2017) provided details on one method on how to create a healthy classroom climate using digital storytelling and spaces for intercultural dialogue.

School ID cards provide a form of ID for students who do not yet have a driver’s license, are ineligible for a State ID card, or do not want to carry their passport around. Regarding ID cards, Shani, a 16-year-old tenth grader from Pakistan, shared,

At school they give you a school identity card, like if a police stops you then you can show the school identification; other is like a school bus on the come back from school. . . . Also in the bus, if they don’t check your identity card, they don’t know if you are from the school or not, you can go to school and get inside; no one is checking you.

Students expressed that they were accustomed to maintaining forms of ID in case they were ever stopped and asked for ID. Similarly, de Graauw (2014) found that in New Haven and San Francisco, municipal ID cards for undocumented immigrants to access basic city services fostered belonging. Belonging also increased when all members of the community have a card, allowing newly-arrived neighbors to “blend in, … stand out less, and it’ll be easier to integrate” (p. 319). Likewise, school-issued ID cards allowed students to identify themselves, access services, feel physically safe, and feel a part of the school community.

Preparing Students for the Future

For students paying some or all of their rent, three additional considerations are as follows: implementing a future orientation curriculum beginning in ninth grade, planning for a two-month welcoming window, and time and budget management support for students simultaneously working a job and attending school. First, some students shared that they did not know their graduation requirements nor did they know the opportunities available to them after high school. Beginning these conversations in the eleventh grade is too late, as their academic profiles and GPAs begin the day they enter high school. Schools should shift quickly from a comfortable, welcoming place to a launch pad to future success with options and awareness, including discussions of individualized plans with students and their families.

Next, schools should consider planning for a suite of welcoming experiences during the first two months of a student’s entry. The students in this study experienced a Newcomer Orientation with their family, personal one-on-one enrollment meetings with their counselors, peer mentors in class (tenth graders), collaborative structures in class, cohorts of students traveling class to class together, and building relationships with peers and teachers. As is typical of newcomer programs, a trickle of new enrollments continues all year, so efforts should be made to track and implement a coherent orientation schedule for each newcomer.

Lastly, students viewed the ability to work while going to school as a huge opportunity. However, most students shared that they never had to balance work and school before and would benefit from explicit support in school. Similarly, in the post-secondary setting, Diaz-Strong et al. (2011) chronicled the incredible financial stress for immigrant students, often ineligible for financial aid for college, and shouldering financial responsibilities of the family’s bills and remittances. Balancing seemingly competing interests of working a job and attending school can cause stress for students (Kreuzer, 2016), and having support in juggling would benefit many.

Implications

Importantly, this research moves the body of literature towards understanding the linkages between newcomer sociocultural experiences and their sense of belonging within secondary newcomer programs. Using a large population of more than 400 high school newcomers from five continents, representing 34 countries, and speaking 21 languages, this study is unique for using previously recognized methods on a significantly larger scale. Further, this study notes five experiential opportunities recommended by newcomers that would impact belonging. Importantly, the data and findings suggest that rather than individual student characteristics or variables predicting school belonging, the experience of being and becoming a part of the new school and developing a feeling of belonging is influenced by social dynamics, relationships, and language. Belonging grows through the actions, practices, and experiences of being a high school student, particularly during times of heightened xenophobia. Schools can apply these findings to build systems, structures, and social relationships to foster belonging for newcomers, paying attention to specific grade-level experiences and English proficiency. The purposeful inclusion of only participants who were high school newcomers allows educators, leaders, and policymakers to learn directly from student voices and ensure schools are a place where newcomers belong.